Birth Order Effects

Helper

source("0_helpers.R")
## Warning: package 'rmarkdown' was built under R version 3.4.3
## Warning: package 'knitr' was built under R version 3.4.3
## 
## Attaching package: 'formr'
## The following object is masked from 'package:rmarkdown':
## 
##     word_document
## Warning: package 'lubridate' was built under R version 3.4.3
## 
## Attaching package: 'lubridate'
## The following object is masked from 'package:base':
## 
##     date
## Warning: package 'stringr' was built under R version 3.4.3
## Loading required package: carData
## lattice theme set by effectsTheme()
## See ?effectsTheme for details.
## 
## Attaching package: 'data.table'
## The following objects are masked from 'package:lubridate':
## 
##     hour, isoweek, mday, minute, month, quarter, second, wday, week, yday, year
## The following objects are masked from 'package:formr':
## 
##     first, last
## Loading required package: Matrix
## 
## Attaching package: 'lmerTest'
## The following object is masked from 'package:lme4':
## 
##     lmer
## The following object is masked from 'package:stats':
## 
##     step
## Warning: package 'haven' was built under R version 3.4.3
## 
## Attaching package: 'psych'
## The following objects are masked from 'package:ggplot2':
## 
##     %+%, alpha
## This is lavaan 0.5-23.1097
## lavaan is BETA software! Please report any bugs.
## 
## Attaching package: 'lavaan'
## The following object is masked from 'package:psych':
## 
##     cor2cov
## Loading required package: lattice
## Loading required package: survival
## Loading required package: Formula
## 
## Attaching package: 'Hmisc'
## The following object is masked from 'package:psych':
## 
##     describe
## The following objects are masked from 'package:base':
## 
##     format.pval, round.POSIXt, trunc.POSIXt, units
## Warning: package 'tidyr' was built under R version 3.4.3
## 
## Attaching package: 'tidyr'
## The following object is masked from 'package:Matrix':
## 
##     expand
## 
## Attaching package: 'dplyr'
## The following objects are masked from 'package:Hmisc':
## 
##     combine, src, summarize
## The following objects are masked from 'package:data.table':
## 
##     between, first, last
## The following objects are masked from 'package:lubridate':
## 
##     intersect, setdiff, union
## The following objects are masked from 'package:formr':
## 
##     first, last
## The following objects are masked from 'package:stats':
## 
##     filter, lag
## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union
## Warning: package 'coefplot' was built under R version 3.4.3
## 
## Attaching package: 'codebook'
## The following object is masked from 'package:psych':
## 
##     bfi
## The following objects are masked from 'package:formr':
## 
##     asis_knit_child, expired, paste.knit_asis
## Warning: package 'effsize' was built under R version 3.4.4
## 
## Attaching package: 'effsize'
## The following object is masked from 'package:psych':
## 
##     cohen.d

Load data

birthorder = readRDS("data/alldata_birthorder.rds")

Data preparations

# For analyses we want to clean the dataset and get rid of all uninteresting data
birthorder = birthorder %>%
  filter(!is.na(pidlink)) %>% # no individuals who are only known from the pregnancy file
  filter(is.na(lifebirths) | lifebirths == 2) %>% # remove 7 and 2 individuals who are known as stillbirth or miscarriage but still have PID
  select(-lifebirths) %>%
  filter(!is.na(mother_pidlink)) %>%
  select(-father_pidlink) %>%
  filter(is.na(any_multiple_birth) | any_multiple_birth != 1) %>% # remove families with twins/triplets/..
  filter(!is.na(birthorder_naive)) %>%
  select(-starts_with("age_"), -wave, -any_multiple_birth, -multiple_birth) %>%
  mutate(money_spent_smoking_log = if_else(is.na(money_spent_smoking_log) & ever_smoked == 0, 0, money_spent_smoking_log),
         amount = if_else(is.na(amount) & ever_smoked == 0, 0, amount),
         amount_still_smokers = if_else(is.na(amount_still_smokers) &  still_smoking == 0, 0, amount_still_smokers),
         birthyear = lubridate::year(birthdate))

# recode Factor Variable as Dummy Variable
birthorder = left_join(birthorder,
                                birthorder %>%
                                  filter(!is.na(Category)) %>%
                                  mutate(var = 1) %>%
                                  select(pidlink, Category, var) %>%
                                  spread(Category, var, fill = 0, sep = "_"), by = "pidlink") %>%
  select(-Category)

# recode Factor Variable as Dummy Variable
birthorder = left_join(birthorder,
                                birthorder %>%
                                  filter(!is.na(Sector)) %>%
                                  mutate(var = 1) %>%
                                  select(pidlink, Sector, var) %>%
                                  spread(Sector, var, fill = 0, sep = "_"), by = "pidlink") %>%
  select(-Sector)

### Variables
birthorder = birthorder %>%
  mutate(
    # center variables that are used for analysis
  g_factor_2015_old = scale(g_factor_2015_old),
  g_factor_2015_young = scale(g_factor_2015_young),
  g_factor_2007_old = scale(g_factor_2007_old),
  g_factor_2007_young = scale(g_factor_2007_young),
  raven_2015_old = scale(raven_2015_old),
  math_2015_old = scale(math_2015_old),
  count_backwards = scale(count_backwards),
  raven_2015_young = scale(raven_2015_young),
  math_2015_young = scale(math_2015_young),
  words_remembered_avg = scale(words_remembered_avg),
  words_immediate = scale(words_immediate),
  words_delayed = scale(words_delayed),
  adaptive_numbering = scale(adaptive_numbering),
  raven_2007_old = scale(raven_2007_old),
  math_2007_old = scale(math_2007_old),
  raven_2007_young = scale(raven_2007_young),
  math_2007_young = scale(math_2007_young),
  riskA = scale(riskA),
  riskB = scale(riskB),
  years_of_education_z = scale(years_of_education),
  Total_score_highest_z = scale(Total_score_highest),
  wage_last_month_z = scale(wage_last_month_log),
  wage_last_year_z = scale(wage_last_year_log),
  big5_ext = scale(big5_ext),
  big5_con = scale(big5_con),
  big5_agree = scale(big5_agree),
  big5_open = scale(big5_open),
  big5_neu = scale(big5_neu),
  attended_school = as.integer(attended_school),
  attended_school = ifelse(attended_school == 1, 0,
                           ifelse(attended_school == 2, 1, NA)))


### Birthorder and Sibling Count
birthorder = birthorder %>% 
  mutate(
# birthorder as factors with levels of 1, 2, 3, 4, 5, 5+
    birthorder_naive_factor = as.character(birthorder_naive),
    birthorder_naive_factor = ifelse(birthorder_naive > 5, "5+",
                                            birthorder_naive_factor),
    birthorder_naive_factor = factor(birthorder_naive_factor, 
                                            levels = c("1","2","3","4","5","5+")),
    sibling_count_naive_factor = as.character(sibling_count_naive),
    sibling_count_naive_factor = ifelse(sibling_count_naive > 5, "5+",
                                               sibling_count_naive_factor),
    sibling_count_naive_factor = factor(sibling_count_naive_factor, 
                                               levels = c("2","3","4","5","5+")),

    birthorder_uterus_alive_factor = as.character(birthorder_uterus_alive),
    birthorder_uterus_alive_factor = ifelse(birthorder_uterus_alive > 5, "5+",
                                            birthorder_uterus_alive_factor),
    birthorder_uterus_alive_factor = factor(birthorder_uterus_alive_factor, 
                                            levels = c("1","2","3","4","5","5+")),
    sibling_count_uterus_alive_factor = as.character(sibling_count_uterus_alive),
    sibling_count_uterus_alive_factor = ifelse(sibling_count_uterus_alive > 5, "5+",
                                               sibling_count_uterus_alive_factor),
    sibling_count_uterus_alive_factor = factor(sibling_count_uterus_alive_factor, 
                                               levels = c("2","3","4","5","5+")),
    birthorder_uterus_preg_factor = as.character(birthorder_uterus_preg),
    birthorder_uterus_preg_factor = ifelse(birthorder_uterus_preg > 5, "5+",
                                           birthorder_uterus_preg_factor),
    birthorder_uterus_preg_factor = factor(birthorder_uterus_preg_factor,
                                           levels = c("1","2","3","4","5","5+")),
    sibling_count_uterus_preg_factor = as.character(sibling_count_uterus_preg),
    sibling_count_uterus_preg_factor = ifelse(sibling_count_uterus_preg > 5, "5+",
                                              sibling_count_uterus_preg_factor),
    sibling_count_uterus_preg_factor = factor(sibling_count_uterus_preg_factor, 
                                              levels = c("2","3","4","5","5+")),
    birthorder_genes_factor = as.character(birthorder_genes),
    birthorder_genes_factor = ifelse(birthorder_genes >5 , "5+", birthorder_genes_factor),
    birthorder_genes_factor = factor(birthorder_genes_factor, 
                                     levels = c("1","2","3","4","5","5+")),
    sibling_count_genes_factor = as.character(sibling_count_genes),
    sibling_count_genes_factor = ifelse(sibling_count_genes >5 , "5+",
                                        sibling_count_genes_factor),
    sibling_count_genes_factor = factor(sibling_count_genes_factor, 
                                        levels = c("2","3","4","5","5+")),
    # interaction birthorder * siblingcout for each birthorder
    count_birthorder_naive =
      factor(str_replace(as.character(interaction(birthorder_naive_factor,                                                              sibling_count_naive_factor)),
                        "\\.", "/"),
                                           levels =   c("1/2","2/2", "1/3",  "2/3",
                                                        "3/3", "1/4", "2/4", "3/4", "4/4",
                                                        "1/5", "2/5", "3/5", "4/5", "5/5",
                                                        "1/5+", "2/5+", "3/5+", "4/5+",
                                                        "5/5+", "5+/5+")),
    count_birthorder_uterus_alive =
      factor(str_replace(as.character(interaction(birthorder_uterus_alive_factor,                                                              sibling_count_uterus_alive_factor)),
                        "\\.", "/"),
                                           levels =   c("1/2","2/2", "1/3",  "2/3",
                                                        "3/3", "1/4", "2/4", "3/4", "4/4",
                                                        "1/5", "2/5", "3/5", "4/5", "5/5",
                                                        "1/5+", "2/5+", "3/5+", "4/5+",
                                                        "5/5+", "5+/5+")),
    count_birthorder_uterus_preg =
      factor(str_replace(as.character(interaction(birthorder_uterus_preg_factor,                                                              sibling_count_uterus_preg_factor)), 
                         "\\.", "/"),
                                           levels =   c("1/2","2/2", "1/3",  "2/3",
                                                        "3/3", "1/4", "2/4", "3/4", "4/4",
                                                        "1/5", "2/5", "3/5", "4/5", "5/5",
                                                        "1/5+", "2/5+", "3/5+", "4/5+",
                                                        "5/5+", "5+/5+")),
    count_birthorder_genes =
      factor(str_replace(as.character(interaction(birthorder_genes_factor,                                                              sibling_count_genes_factor)), "\\.", "/"),
                                           levels =   c("1/2","2/2", "1/3",  "2/3",
                                                        "3/3", "1/4", "2/4", "3/4", "4/4",
                                                        "1/5", "2/5", "3/5", "4/5", "5/5",
                                                        "1/5+", "2/5+", "3/5+", "4/5+",
                                                        "5/5+", "5+/5+")))

birthorder <- birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive)

Intelligence

g-factor 2015 old

birthorder <- birthorder %>% mutate(outcome = g_factor_2015_old)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count +
               (1 | mother_pidlink),
                   data = birthorder)


compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.1911 0.2 0.9555 -0.2009 0.583 fixed
poly(age, 3, raw = TRUE)1 0.0243 0.02056 1.182 -0.01598 0.06459 fixed
poly(age, 3, raw = TRUE)2 -0.001017 0.0006584 -1.544 -0.002307 0.0002738 fixed
poly(age, 3, raw = TRUE)3 0.000003728 0.000006632 0.5621 -0.000009271 0.00001673 fixed
male 0.06205 0.01493 4.156 0.03279 0.09131 fixed
sibling_count3 0.03391 0.03437 0.9865 -0.03346 0.1013 fixed
sibling_count4 -0.0006889 0.03585 -0.01922 -0.07095 0.06957 fixed
sibling_count5 -0.001687 0.03755 -0.04492 -0.07528 0.07191 fixed
sibling_count5+ -0.1915 0.02919 -6.56 -0.2487 -0.1343 fixed
sd_(Intercept).mother_pidlink 0.5824 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7438 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.1824 0.2002 0.9114 -0.2099 0.5748 fixed
birth_order -0.003077 0.00332 -0.9269 -0.009583 0.003429 fixed
poly(age, 3, raw = TRUE)1 0.02632 0.02067 1.273 -0.0142 0.06684 fixed
poly(age, 3, raw = TRUE)2 -0.001092 0.0006635 -1.646 -0.002393 0.0002085 fixed
poly(age, 3, raw = TRUE)3 0.00000448 0.000006683 0.6704 -0.000008618 0.00001758 fixed
male 0.06218 0.01493 4.164 0.03292 0.09145 fixed
sibling_count3 0.03436 0.03437 0.9997 -0.033 0.1017 fixed
sibling_count4 0.001231 0.0359 0.0343 -0.06913 0.07159 fixed
sibling_count5 0.001859 0.03773 0.04927 -0.07209 0.07581 fixed
sibling_count5+ -0.18 0.0317 -5.679 -0.2422 -0.1179 fixed
sd_(Intercept).mother_pidlink 0.582 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.744 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.1974 0.2017 0.9789 -0.1979 0.5928 fixed
poly(age, 3, raw = TRUE)1 0.02456 0.02074 1.184 -0.01608 0.0652 fixed
poly(age, 3, raw = TRUE)2 -0.001027 0.0006661 -1.542 -0.002332 0.0002787 fixed
poly(age, 3, raw = TRUE)3 0.000003763 0.000006717 0.5601 -0.000009403 0.00001693 fixed
male 0.06205 0.01493 4.155 0.03278 0.09131 fixed
sibling_count3 0.03503 0.0347 1.01 -0.03298 0.103 fixed
sibling_count4 0.005935 0.03659 0.1622 -0.06578 0.07765 fixed
sibling_count5 0.001283 0.03871 0.03314 -0.07459 0.07716 fixed
sibling_count5+ -0.1803 0.03284 -5.489 -0.2446 -0.1159 fixed
birth_order_nonlinear2 -0.01433 0.02147 -0.6675 -0.05641 0.02775 fixed
birth_order_nonlinear3 -0.0118 0.02501 -0.4716 -0.06082 0.03723 fixed
birth_order_nonlinear4 -0.03464 0.0284 -1.22 -0.0903 0.02102 fixed
birth_order_nonlinear5 0.01513 0.03228 0.4688 -0.04814 0.07841 fixed
birth_order_nonlinear5+ -0.02847 0.02793 -1.019 -0.08321 0.02627 fixed
sd_(Intercept).mother_pidlink 0.5822 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.744 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.1955 0.203 0.9634 -0.2023 0.5933 fixed
poly(age, 3, raw = TRUE)1 0.02567 0.02079 1.235 -0.01507 0.06642 fixed
poly(age, 3, raw = TRUE)2 -0.00106 0.0006683 -1.586 -0.00237 0.00025 fixed
poly(age, 3, raw = TRUE)3 0.000004077 0.000006744 0.6046 -0.00000914 0.0000173 fixed
male 0.06191 0.01493 4.145 0.03264 0.09118 fixed
count_birth_order2/2 -0.04118 0.04236 -0.9722 -0.1242 0.04184 fixed
count_birth_order1/3 0.02571 0.04281 0.6006 -0.0582 0.1096 fixed
count_birth_order2/3 0.01942 0.04732 0.4103 -0.07334 0.1122 fixed
count_birth_order3/3 0.0007667 0.05251 0.0146 -0.1021 0.1037 fixed
count_birth_order1/4 -0.01819 0.04862 -0.374 -0.1135 0.07711 fixed
count_birth_order2/4 0.01964 0.05083 0.3863 -0.08 0.1193 fixed
count_birth_order3/4 -0.02461 0.05444 -0.4519 -0.1313 0.0821 fixed
count_birth_order4/4 -0.059 0.05719 -1.032 -0.1711 0.05308 fixed
count_birth_order1/5 -0.09978 0.05416 -1.842 -0.2059 0.006367 fixed
count_birth_order2/5 -0.00005492 0.05667 -0.000969 -0.1111 0.111 fixed
count_birth_order3/5 0.003105 0.05811 0.05344 -0.1108 0.117 fixed
count_birth_order4/5 -0.01033 0.06116 -0.1689 -0.1302 0.1095 fixed
count_birth_order5/5 0.0591 0.06229 0.9489 -0.06298 0.1812 fixed
count_birth_order1/5+ -0.1474 0.04374 -3.371 -0.2332 -0.06171 fixed
count_birth_order2/5+ -0.2332 0.04485 -5.198 -0.3211 -0.1452 fixed
count_birth_order3/5+ -0.1994 0.04399 -4.531 -0.2856 -0.1131 fixed
count_birth_order4/5+ -0.2272 0.04306 -5.276 -0.3116 -0.1428 fixed
count_birth_order5/5+ -0.1914 0.04332 -4.419 -0.2764 -0.1065 fixed
count_birth_order5+/5+ -0.2187 0.03572 -6.122 -0.2887 -0.1487 fixed
sd_(Intercept).mother_pidlink 0.5821 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7439 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 35518 35600 -17748 35496 NA NA NA
12 35519 35609 -17747 35495 0.8602 1 0.3537
16 35524 35644 -17746 35492 2.576 4 0.631
26 35532 35728 -17740 35480 11.82 10 0.2971

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.8339 0.3708 -2.249 -1.561 -0.1072 fixed
poly(age, 3, raw = TRUE)1 0.1467 0.04203 3.489 0.06427 0.229 fixed
poly(age, 3, raw = TRUE)2 -0.004903 0.001499 -3.271 -0.007841 -0.001965 fixed
poly(age, 3, raw = TRUE)3 0.00004724 0.00001697 2.784 0.00001399 0.0000805 fixed
male -0.02065 0.02158 -0.9572 -0.06294 0.02164 fixed
sibling_count3 0.0003067 0.03701 0.008289 -0.07222 0.07284 fixed
sibling_count4 -0.0771 0.04036 -1.91 -0.1562 0.002002 fixed
sibling_count5 -0.1511 0.04636 -3.26 -0.242 -0.06025 fixed
sibling_count5+ -0.2821 0.04094 -6.891 -0.3624 -0.2019 fixed
sd_(Intercept).mother_pidlink 0.5177 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.6977 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.8412 0.3708 -2.269 -1.568 -0.1145 fixed
birth_order 0.009041 0.007302 1.238 -0.005271 0.02335 fixed
poly(age, 3, raw = TRUE)1 0.146 0.04203 3.473 0.0636 0.2284 fixed
poly(age, 3, raw = TRUE)2 -0.004888 0.001499 -3.261 -0.007826 -0.00195 fixed
poly(age, 3, raw = TRUE)3 0.00004746 0.00001697 2.797 0.0000142 0.00008071 fixed
male -0.02114 0.02158 -0.9795 -0.06343 0.02116 fixed
sibling_count3 -0.004319 0.0372 -0.1161 -0.07722 0.06858 fixed
sibling_count4 -0.08824 0.04135 -2.134 -0.1693 -0.007188 fixed
sibling_count5 -0.169 0.04857 -3.48 -0.2642 -0.07383 fixed
sibling_count5+ -0.3178 0.05005 -6.349 -0.4159 -0.2197 fixed
sd_(Intercept).mother_pidlink 0.518 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.6974 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.8705 0.3718 -2.341 -1.599 -0.1418 fixed
poly(age, 3, raw = TRUE)1 0.1492 0.04209 3.546 0.06675 0.2317 fixed
poly(age, 3, raw = TRUE)2 -0.005006 0.001501 -3.336 -0.007947 -0.002064 fixed
poly(age, 3, raw = TRUE)3 0.00004879 0.00001699 2.872 0.0000155 0.00008209 fixed
male -0.02051 0.02158 -0.9504 -0.06281 0.02179 fixed
sibling_count3 -0.006197 0.0378 -0.164 -0.08028 0.06788 fixed
sibling_count4 -0.08533 0.04264 -2.001 -0.1689 -0.001759 fixed
sibling_count5 -0.175 0.05049 -3.466 -0.2739 -0.07602 fixed
sibling_count5+ -0.3206 0.05119 -6.263 -0.421 -0.2203 fixed
birth_order_nonlinear2 0.03959 0.02714 1.459 -0.01361 0.09279 fixed
birth_order_nonlinear3 0.02682 0.03361 0.7981 -0.03905 0.09269 fixed
birth_order_nonlinear4 0.0122 0.04178 0.292 -0.06969 0.09409 fixed
birth_order_nonlinear5 0.09656 0.05185 1.862 -0.005073 0.1982 fixed
birth_order_nonlinear5+ 0.06261 0.05372 1.166 -0.04268 0.1679 fixed
sd_(Intercept).mother_pidlink 0.5179 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.6975 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.8752 0.3728 -2.348 -1.606 -0.1446 fixed
poly(age, 3, raw = TRUE)1 0.1489 0.04219 3.53 0.06625 0.2316 fixed
poly(age, 3, raw = TRUE)2 -0.004981 0.001505 -3.311 -0.00793 -0.002032 fixed
poly(age, 3, raw = TRUE)3 0.00004837 0.00001704 2.839 0.00001498 0.00008176 fixed
male -0.02137 0.0216 -0.9894 -0.06371 0.02097 fixed
count_birth_order2/2 0.05487 0.04968 1.105 -0.0425 0.1522 fixed
count_birth_order1/3 -0.006477 0.04639 -0.1396 -0.0974 0.08444 fixed
count_birth_order2/3 0.04298 0.05006 0.8586 -0.05513 0.1411 fixed
count_birth_order3/3 0.02919 0.0557 0.5241 -0.07999 0.1384 fixed
count_birth_order1/4 -0.1094 0.05641 -1.939 -0.22 0.001176 fixed
count_birth_order2/4 -0.02503 0.05803 -0.4314 -0.1388 0.08871 fixed
count_birth_order3/4 -0.06345 0.06065 -1.046 -0.1823 0.05541 fixed
count_birth_order4/4 -0.03422 0.06329 -0.5408 -0.1583 0.08982 fixed
count_birth_order1/5 -0.1232 0.07517 -1.639 -0.2706 0.02409 fixed
count_birth_order2/5 -0.0938 0.08081 -1.161 -0.2522 0.06458 fixed
count_birth_order3/5 -0.1584 0.07568 -2.093 -0.3067 -0.01005 fixed
count_birth_order4/5 -0.1877 0.07323 -2.563 -0.3312 -0.04416 fixed
count_birth_order5/5 -0.1089 0.07511 -1.45 -0.2561 0.03829 fixed
count_birth_order1/5+ -0.2572 0.07465 -3.446 -0.4035 -0.1109 fixed
count_birth_order2/5+ -0.373 0.07449 -5.008 -0.519 -0.227 fixed
count_birth_order3/5+ -0.2644 0.07378 -3.583 -0.409 -0.1198 fixed
count_birth_order4/5+ -0.3211 0.06919 -4.641 -0.4567 -0.1855 fixed
count_birth_order5/5+ -0.1957 0.06586 -2.972 -0.3248 -0.06663 fixed
count_birth_order5+/5+ -0.2531 0.05163 -4.903 -0.3543 -0.152 fixed
sd_(Intercept).mother_pidlink 0.5182 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.6977 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 14535 14608 -7257 14513 NA NA NA
12 14536 14616 -7256 14512 1.533 1 0.2156
16 14540 14647 -7254 14508 3.603 4 0.4623
26 14553 14727 -7251 14501 6.45 10 0.7761

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.8021 0.3701 -2.168 -1.527 -0.07681 fixed
poly(age, 3, raw = TRUE)1 0.1431 0.04197 3.41 0.06086 0.2254 fixed
poly(age, 3, raw = TRUE)2 -0.004819 0.001497 -3.219 -0.007753 -0.001885 fixed
poly(age, 3, raw = TRUE)3 0.00004624 0.00001695 2.728 0.00001302 0.00007946 fixed
male -0.02123 0.02151 -0.9869 -0.0634 0.02093 fixed
sibling_count3 0.006856 0.03997 0.1715 -0.07149 0.0852 fixed
sibling_count4 -0.0426 0.0426 -1 -0.1261 0.04089 fixed
sibling_count5 -0.07804 0.04578 -1.705 -0.1678 0.01168 fixed
sibling_count5+ -0.1916 0.04007 -4.783 -0.2702 -0.1131 fixed
sd_(Intercept).mother_pidlink 0.5208 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.6974 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.7996 0.3702 -2.16 -1.525 -0.07407 fixed
birth_order -0.002276 0.006431 -0.354 -0.01488 0.01033 fixed
poly(age, 3, raw = TRUE)1 0.1432 0.04197 3.412 0.06094 0.2255 fixed
poly(age, 3, raw = TRUE)2 -0.00482 0.001497 -3.219 -0.007754 -0.001885 fixed
poly(age, 3, raw = TRUE)3 0.00004615 0.00001695 2.722 0.00001293 0.00007938 fixed
male -0.02113 0.02152 -0.9817 -0.0633 0.02105 fixed
sibling_count3 0.008026 0.04011 0.2001 -0.07059 0.08664 fixed
sibling_count4 -0.03991 0.04327 -0.9224 -0.1247 0.0449 fixed
sibling_count5 -0.07387 0.04727 -1.563 -0.1665 0.01878 fixed
sibling_count5+ -0.183 0.04692 -3.9 -0.275 -0.09105 fixed
sd_(Intercept).mother_pidlink 0.5206 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.6975 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.8363 0.371 -2.254 -1.564 -0.1091 fixed
poly(age, 3, raw = TRUE)1 0.1458 0.04202 3.469 0.06341 0.2281 fixed
poly(age, 3, raw = TRUE)2 -0.004913 0.001499 -3.278 -0.00785 -0.001975 fixed
poly(age, 3, raw = TRUE)3 0.00004722 0.00001697 2.782 0.00001395 0.00008048 fixed
male -0.02023 0.02152 -0.9401 -0.06242 0.02195 fixed
sibling_count3 0.008151 0.04069 0.2003 -0.07159 0.0879 fixed
sibling_count4 -0.03462 0.04449 -0.7782 -0.1218 0.05258 fixed
sibling_count5 -0.0762 0.04905 -1.554 -0.1723 0.01993 fixed
sibling_count5+ -0.1822 0.04805 -3.792 -0.2764 -0.08804 fixed
birth_order_nonlinear2 0.0345 0.02777 1.242 -0.01993 0.08893 fixed
birth_order_nonlinear3 -0.005354 0.03358 -0.1595 -0.07117 0.06046 fixed
birth_order_nonlinear4 -0.02399 0.0407 -0.5894 -0.1038 0.05579 fixed
birth_order_nonlinear5 0.04948 0.0496 0.9975 -0.04774 0.1467 fixed
birth_order_nonlinear5+ -0.009693 0.04843 -0.2002 -0.1046 0.08522 fixed
sd_(Intercept).mother_pidlink 0.5206 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.6975 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.8476 0.3718 -2.28 -1.576 -0.1188 fixed
poly(age, 3, raw = TRUE)1 0.1447 0.04207 3.438 0.0622 0.2271 fixed
poly(age, 3, raw = TRUE)2 -0.004861 0.001501 -3.238 -0.007803 -0.001919 fixed
poly(age, 3, raw = TRUE)3 0.00004649 0.000017 2.734 0.00001317 0.00007981 fixed
male -0.02115 0.02153 -0.9822 -0.06334 0.02105 fixed
count_birth_order2/2 0.09269 0.05458 1.698 -0.01428 0.1997 fixed
count_birth_order1/3 0.03991 0.05027 0.7939 -0.05862 0.1384 fixed
count_birth_order2/3 0.02977 0.05389 0.5524 -0.07585 0.1354 fixed
count_birth_order3/3 0.04257 0.06031 0.7058 -0.07564 0.1608 fixed
count_birth_order1/4 -0.07608 0.05917 -1.286 -0.1921 0.03989 fixed
count_birth_order2/4 0.08764 0.06018 1.456 -0.03032 0.2056 fixed
count_birth_order3/4 -0.0303 0.065 -0.4661 -0.1577 0.0971 fixed
count_birth_order4/4 -0.03003 0.06725 -0.4466 -0.1618 0.1018 fixed
count_birth_order1/5 -0.0001437 0.06919 -0.002077 -0.1357 0.1355 fixed
count_birth_order2/5 -0.01429 0.07426 -0.1925 -0.1598 0.1313 fixed
count_birth_order3/5 -0.1014 0.07188 -1.411 -0.2423 0.03948 fixed
count_birth_order4/5 -0.1037 0.07452 -1.391 -0.2497 0.04236 fixed
count_birth_order5/5 -0.02898 0.07439 -0.3896 -0.1748 0.1168 fixed
count_birth_order1/5+ -0.1089 0.06599 -1.65 -0.2382 0.02047 fixed
count_birth_order2/5+ -0.2303 0.06883 -3.346 -0.3652 -0.09542 fixed
count_birth_order3/5+ -0.1537 0.06699 -2.294 -0.285 -0.02239 fixed
count_birth_order4/5+ -0.1817 0.06466 -2.81 -0.3084 -0.05497 fixed
count_birth_order5/5+ -0.1002 0.06589 -1.52 -0.2293 0.02896 fixed
count_birth_order5+/5+ -0.1736 0.05057 -3.433 -0.2727 -0.07449 fixed
sd_(Intercept).mother_pidlink 0.5209 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.6971 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 14682 14755 -7330 14660 NA NA NA
12 14684 14764 -7330 14660 0.1261 1 0.7225
16 14687 14794 -7328 14655 4.215 4 0.3777
26 14694 14868 -7321 14642 13.01 10 0.2232

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.9226 0.3751 -2.46 -1.658 -0.1875 fixed
poly(age, 3, raw = TRUE)1 0.1559 0.04254 3.665 0.07254 0.2393 fixed
poly(age, 3, raw = TRUE)2 -0.005258 0.001517 -3.465 -0.008231 -0.002284 fixed
poly(age, 3, raw = TRUE)3 0.0000515 0.00001718 2.997 0.00001782 0.00008518 fixed
male -0.02117 0.0218 -0.9713 -0.06389 0.02155 fixed
sibling_count3 0.01472 0.03647 0.4035 -0.05677 0.0862 fixed
sibling_count4 -0.06404 0.04 -1.601 -0.1424 0.01437 fixed
sibling_count5 -0.1279 0.04754 -2.691 -0.2211 -0.03476 fixed
sibling_count5+ -0.271 0.04142 -6.541 -0.3522 -0.1898 fixed
sd_(Intercept).mother_pidlink 0.5155 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.6986 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.9298 0.3751 -2.479 -1.665 -0.1947 fixed
birth_order 0.01005 0.007512 1.338 -0.004672 0.02477 fixed
poly(age, 3, raw = TRUE)1 0.1551 0.04254 3.646 0.07172 0.2385 fixed
poly(age, 3, raw = TRUE)2 -0.005241 0.001517 -3.454 -0.008214 -0.002267 fixed
poly(age, 3, raw = TRUE)3 0.00005174 0.00001718 3.011 0.00001806 0.00008542 fixed
male -0.0215 0.0218 -0.9866 -0.06422 0.02122 fixed
sibling_count3 0.009523 0.03668 0.2596 -0.06237 0.08141 fixed
sibling_count4 -0.07619 0.04102 -1.857 -0.1566 0.004212 fixed
sibling_count5 -0.147 0.04963 -2.962 -0.2443 -0.04973 fixed
sibling_count5+ -0.3101 0.0507 -6.116 -0.4095 -0.2107 fixed
sd_(Intercept).mother_pidlink 0.5157 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.6984 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.9568 0.3761 -2.544 -1.694 -0.2197 fixed
poly(age, 3, raw = TRUE)1 0.1581 0.0426 3.711 0.07459 0.2416 fixed
poly(age, 3, raw = TRUE)2 -0.005346 0.001519 -3.519 -0.008323 -0.002369 fixed
poly(age, 3, raw = TRUE)3 0.0000529 0.0000172 3.075 0.00001918 0.00008662 fixed
male -0.0212 0.0218 -0.9723 -0.06393 0.02153 fixed
sibling_count3 0.00883 0.03729 0.2368 -0.06427 0.08192 fixed
sibling_count4 -0.07612 0.04236 -1.797 -0.1591 0.006895 fixed
sibling_count5 -0.1499 0.05142 -2.914 -0.2506 -0.04907 fixed
sibling_count5+ -0.3102 0.05189 -5.978 -0.4119 -0.2085 fixed
birth_order_nonlinear2 0.04352 0.02709 1.606 -0.00958 0.09662 fixed
birth_order_nonlinear3 0.02617 0.03362 0.7785 -0.03972 0.09207 fixed
birth_order_nonlinear4 0.0344 0.04292 0.8015 -0.04972 0.1185 fixed
birth_order_nonlinear5 0.07337 0.05415 1.355 -0.03276 0.1795 fixed
birth_order_nonlinear5+ 0.06926 0.0553 1.252 -0.03912 0.1776 fixed
sd_(Intercept).mother_pidlink 0.5155 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.6987 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.9636 0.3773 -2.554 -1.703 -0.2242 fixed
poly(age, 3, raw = TRUE)1 0.1579 0.04271 3.697 0.07418 0.2416 fixed
poly(age, 3, raw = TRUE)2 -0.005325 0.001524 -3.495 -0.008311 -0.002339 fixed
poly(age, 3, raw = TRUE)3 0.0000525 0.00001726 3.042 0.00001867 0.00008633 fixed
male -0.02235 0.02182 -1.024 -0.06513 0.02042 fixed
count_birth_order2/2 0.06086 0.04827 1.261 -0.03375 0.1555 fixed
count_birth_order1/3 0.007247 0.04578 0.1583 -0.08247 0.09696 fixed
count_birth_order2/3 0.0644 0.04994 1.29 -0.03348 0.1623 fixed
count_birth_order3/3 0.04643 0.05471 0.8486 -0.0608 0.1537 fixed
count_birth_order1/4 -0.09561 0.05654 -1.691 -0.2064 0.0152 fixed
count_birth_order2/4 -0.0208 0.05792 -0.3592 -0.1343 0.09272 fixed
count_birth_order3/4 -0.04594 0.06005 -0.765 -0.1636 0.07176 fixed
count_birth_order4/4 -0.004906 0.06335 -0.07745 -0.1291 0.1192 fixed
count_birth_order1/5 -0.1047 0.07511 -1.394 -0.252 0.04248 fixed
count_birth_order2/5 -0.07203 0.08321 -0.8656 -0.2351 0.09106 fixed
count_birth_order3/5 -0.1374 0.0791 -1.738 -0.2925 0.01759 fixed
count_birth_order4/5 -0.1243 0.0766 -1.623 -0.2745 0.02578 fixed
count_birth_order5/5 -0.112 0.08027 -1.395 -0.2693 0.04534 fixed
count_birth_order1/5+ -0.2309 0.07637 -3.024 -0.3806 -0.08123 fixed
count_birth_order2/5+ -0.3469 0.07643 -4.539 -0.4967 -0.1971 fixed
count_birth_order3/5+ -0.2686 0.07462 -3.599 -0.4149 -0.1223 fixed
count_birth_order4/5+ -0.3014 0.07262 -4.15 -0.4437 -0.1591 fixed
count_birth_order5/5+ -0.2063 0.06741 -3.061 -0.3385 -0.07422 fixed
count_birth_order5+/5+ -0.2356 0.05257 -4.482 -0.3386 -0.1326 fixed
sd_(Intercept).mother_pidlink 0.5157 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.699 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 14247 14320 -7112 14225 NA NA NA
12 14247 14327 -7111 14223 1.792 1 0.1807
16 14253 14359 -7110 14221 2.106 4 0.7163
26 14267 14440 -7108 14215 5.623 10 0.8458

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

g-factor 2015 young

birthorder <- birthorder %>% mutate(outcome = g_factor_2015_young)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count +
               (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.216 0.1156 -27.82 -3.443 -2.99 fixed
poly(age, 3, raw = TRUE)1 0.4938 0.02146 23.01 0.4517 0.5358 fixed
poly(age, 3, raw = TRUE)2 -0.02095 0.001245 -16.83 -0.02339 -0.01851 fixed
poly(age, 3, raw = TRUE)3 0.0002754 0.00002287 12.04 0.0002305 0.0003202 fixed
male -0.06654 0.01647 -4.04 -0.09882 -0.03426 fixed
sibling_count3 -0.02258 0.02855 -0.791 -0.07853 0.03337 fixed
sibling_count4 -0.1055 0.03201 -3.295 -0.1682 -0.04272 fixed
sibling_count5 -0.07458 0.03622 -2.059 -0.1456 -0.003602 fixed
sibling_count5+ -0.2683 0.02871 -9.345 -0.3245 -0.212 fixed
sd_(Intercept).mother_pidlink 0.5086 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7822 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.201 0.116 -27.6 -3.428 -2.974 fixed
birth_order -0.006933 0.004416 -1.57 -0.01559 0.001723 fixed
poly(age, 3, raw = TRUE)1 0.4933 0.02146 22.98 0.4512 0.5353 fixed
poly(age, 3, raw = TRUE)2 -0.02094 0.001245 -16.82 -0.02338 -0.0185 fixed
poly(age, 3, raw = TRUE)3 0.0002751 0.00002287 12.03 0.0002303 0.0003199 fixed
male -0.06643 0.01647 -4.033 -0.09871 -0.03415 fixed
sibling_count3 -0.01848 0.02866 -0.6447 -0.07465 0.0377 fixed
sibling_count4 -0.0956 0.03262 -2.931 -0.1595 -0.03167 fixed
sibling_count5 -0.05868 0.0376 -1.561 -0.1324 0.01502 fixed
sibling_count5+ -0.2291 0.03801 -6.028 -0.3036 -0.1546 fixed
sd_(Intercept).mother_pidlink 0.5082 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7824 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.222 0.1171 -27.51 -3.452 -2.993 fixed
poly(age, 3, raw = TRUE)1 0.4949 0.02154 22.98 0.4527 0.5371 fixed
poly(age, 3, raw = TRUE)2 -0.02102 0.001247 -16.86 -0.02346 -0.01857 fixed
poly(age, 3, raw = TRUE)3 0.0002761 0.00002288 12.07 0.0002313 0.000321 fixed
male -0.06671 0.01647 -4.051 -0.09899 -0.03443 fixed
sibling_count3 -0.03268 0.02952 -1.107 -0.09054 0.02518 fixed
sibling_count4 -0.1082 0.03513 -3.081 -0.1771 -0.03938 fixed
sibling_count5 -0.0603 0.04138 -1.457 -0.1414 0.02081 fixed
sibling_count5+ -0.2275 0.04184 -5.437 -0.3095 -0.1455 fixed
birth_order_nonlinear2 0.006876 0.02242 0.3068 -0.03706 0.05081 fixed
birth_order_nonlinear3 0.0427 0.02829 1.509 -0.01275 0.09815 fixed
birth_order_nonlinear4 -0.03406 0.03506 -0.9713 -0.1028 0.03467 fixed
birth_order_nonlinear5 -0.04478 0.04123 -1.086 -0.1256 0.03602 fixed
birth_order_nonlinear5+ -0.05059 0.03926 -1.289 -0.1275 0.02635 fixed
sd_(Intercept).mother_pidlink 0.509 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7819 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.234 0.1181 -27.37 -3.465 -3.002 fixed
poly(age, 3, raw = TRUE)1 0.4962 0.02158 22.99 0.4539 0.5385 fixed
poly(age, 3, raw = TRUE)2 -0.0211 0.001249 -16.9 -0.02354 -0.01865 fixed
poly(age, 3, raw = TRUE)3 0.0002775 0.00002291 12.11 0.0002326 0.0003224 fixed
male -0.06686 0.01647 -4.058 -0.09915 -0.03457 fixed
count_birth_order2/2 0.02197 0.03654 0.6012 -0.04965 0.09358 fixed
count_birth_order1/3 -0.03817 0.03737 -1.021 -0.1114 0.03507 fixed
count_birth_order2/3 -0.01761 0.0381 -0.4622 -0.09229 0.05707 fixed
count_birth_order3/3 0.02761 0.04314 0.64 -0.05694 0.1122 fixed
count_birth_order1/4 -0.103 0.05108 -2.017 -0.2031 -0.002897 fixed
count_birth_order2/4 -0.07822 0.04853 -1.612 -0.1733 0.01689 fixed
count_birth_order3/4 -0.1103 0.04572 -2.412 -0.1999 -0.02068 fixed
count_birth_order4/4 -0.0947 0.0499 -1.898 -0.1925 0.003108 fixed
count_birth_order1/5 -0.01787 0.0693 -0.2578 -0.1537 0.118 fixed
count_birth_order2/5 -0.0782 0.06596 -1.186 -0.2075 0.05109 fixed
count_birth_order3/5 0.01626 0.05977 0.272 -0.1009 0.1334 fixed
count_birth_order4/5 -0.1432 0.05618 -2.549 -0.2533 -0.0331 fixed
count_birth_order5/5 -0.07428 0.05573 -1.333 -0.1835 0.03494 fixed
count_birth_order1/5+ -0.1652 0.06939 -2.38 -0.3012 -0.02916 fixed
count_birth_order2/5+ -0.2743 0.06498 -4.221 -0.4017 -0.1469 fixed
count_birth_order3/5+ -0.1406 0.05784 -2.431 -0.254 -0.02727 fixed
count_birth_order4/5+ -0.2586 0.05264 -4.912 -0.3618 -0.1554 fixed
count_birth_order5/5+ -0.2829 0.04729 -5.982 -0.3756 -0.1902 fixed
count_birth_order5+/5+ -0.2731 0.03302 -8.271 -0.3379 -0.2084 fixed
sd_(Intercept).mother_pidlink 0.5085 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7822 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 31446 31527 -15712 31424 NA NA NA
12 31446 31534 -15711 31422 2.467 1 0.1162
16 31448 31566 -15708 31416 6.164 4 0.1872
26 31459 31651 -15703 31407 8.923 10 0.5394

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.724 0.1782 -20.9 -4.074 -3.375 fixed
poly(age, 3, raw = TRUE)1 0.5997 0.03658 16.4 0.528 0.6714 fixed
poly(age, 3, raw = TRUE)2 -0.02751 0.002318 -11.87 -0.03205 -0.02297 fixed
poly(age, 3, raw = TRUE)3 0.0004057 0.00004565 8.887 0.0003162 0.0004952 fixed
male -0.08654 0.01931 -4.481 -0.1244 -0.04869 fixed
sibling_count3 -0.03864 0.02747 -1.407 -0.09248 0.0152 fixed
sibling_count4 -0.1044 0.03372 -3.095 -0.1705 -0.03828 fixed
sibling_count5 -0.2253 0.04329 -5.203 -0.3101 -0.1404 fixed
sibling_count5+ -0.2776 0.04026 -6.896 -0.3565 -0.1987 fixed
sd_(Intercept).mother_pidlink 0.5011 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.772 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.711 0.1805 -20.55 -4.064 -3.357 fixed
birth_order -0.004161 0.008678 -0.4795 -0.02117 0.01285 fixed
poly(age, 3, raw = TRUE)1 0.5985 0.03668 16.32 0.5266 0.6704 fixed
poly(age, 3, raw = TRUE)2 -0.02745 0.002322 -11.82 -0.032 -0.0229 fixed
poly(age, 3, raw = TRUE)3 0.0004046 0.00004572 8.85 0.000315 0.0004942 fixed
male -0.08651 0.01931 -4.479 -0.1244 -0.04865 fixed
sibling_count3 -0.03584 0.02808 -1.276 -0.09088 0.01921 fixed
sibling_count4 -0.09815 0.03614 -2.716 -0.169 -0.02731 fixed
sibling_count5 -0.2149 0.04835 -4.445 -0.3097 -0.1202 fixed
sibling_count5+ -0.2575 0.05816 -4.427 -0.3715 -0.1435 fixed
sd_(Intercept).mother_pidlink 0.5009 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7721 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.74 0.1801 -20.77 -4.093 -3.387 fixed
poly(age, 3, raw = TRUE)1 0.6015 0.03669 16.39 0.5296 0.6734 fixed
poly(age, 3, raw = TRUE)2 -0.02762 0.002322 -11.9 -0.03217 -0.02307 fixed
poly(age, 3, raw = TRUE)3 0.0004079 0.00004571 8.922 0.0003183 0.0004975 fixed
male -0.08719 0.01931 -4.514 -0.125 -0.04934 fixed
sibling_count3 -0.04374 0.0293 -1.493 -0.1012 0.01368 fixed
sibling_count4 -0.1057 0.03893 -2.716 -0.182 -0.02943 fixed
sibling_count5 -0.1916 0.05298 -3.616 -0.2954 -0.08775 fixed
sibling_count5+ -0.2788 0.06273 -4.444 -0.4017 -0.1559 fixed
birth_order_nonlinear2 0.02003 0.0231 0.867 -0.02525 0.06531 fixed
birth_order_nonlinear3 0.01649 0.03236 0.5095 -0.04694 0.07993 fixed
birth_order_nonlinear4 -0.003464 0.04434 -0.07813 -0.09037 0.08344 fixed
birth_order_nonlinear5 -0.09202 0.05964 -1.543 -0.2089 0.02488 fixed
birth_order_nonlinear5+ 0.03902 0.06915 0.5643 -0.09651 0.1745 fixed
sd_(Intercept).mother_pidlink 0.5014 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7717 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.749 0.181 -20.72 -4.104 -3.395 fixed
poly(age, 3, raw = TRUE)1 0.6024 0.03679 16.37 0.5303 0.6745 fixed
poly(age, 3, raw = TRUE)2 -0.02769 0.002328 -11.89 -0.03225 -0.02313 fixed
poly(age, 3, raw = TRUE)3 0.0004095 0.00004585 8.932 0.0003196 0.0004994 fixed
male -0.08649 0.01933 -4.474 -0.1244 -0.0486 fixed
count_birth_order2/2 0.03737 0.03244 1.152 -0.02622 0.101 fixed
count_birth_order1/3 -0.03075 0.03821 -0.8046 -0.1056 0.04415 fixed
count_birth_order2/3 -0.02697 0.03676 -0.7337 -0.09902 0.04507 fixed
count_birth_order3/3 -0.01657 0.04023 -0.4119 -0.09543 0.06229 fixed
count_birth_order1/4 -0.08078 0.06022 -1.341 -0.1988 0.03726 fixed
count_birth_order2/4 -0.09773 0.05287 -1.849 -0.2014 0.005892 fixed
count_birth_order3/4 -0.118 0.05048 -2.337 -0.2169 -0.01903 fixed
count_birth_order4/4 -0.06776 0.05001 -1.355 -0.1658 0.03025 fixed
count_birth_order1/5 -0.1764 0.1048 -1.683 -0.3818 0.02899 fixed
count_birth_order2/5 -0.125 0.09341 -1.338 -0.3081 0.05812 fixed
count_birth_order3/5 -0.09885 0.07912 -1.249 -0.2539 0.05621 fixed
count_birth_order4/5 -0.2459 0.06553 -3.753 -0.3744 -0.1175 fixed
count_birth_order5/5 -0.2798 0.0636 -4.399 -0.4044 -0.1551 fixed
count_birth_order1/5+ -0.2669 0.1339 -1.994 -0.5293 -0.004587 fixed
count_birth_order2/5+ -0.2542 0.1225 -2.076 -0.4943 -0.01421 fixed
count_birth_order3/5+ -0.245 0.1013 -2.418 -0.4436 -0.04641 fixed
count_birth_order4/5+ -0.2884 0.09095 -3.171 -0.4667 -0.1101 fixed
count_birth_order5/5+ -0.3627 0.07314 -4.959 -0.5061 -0.2194 fixed
count_birth_order5+/5+ -0.2334 0.04879 -4.784 -0.329 -0.1378 fixed
sd_(Intercept).mother_pidlink 0.5007 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7724 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 22331 22408 -11154 22309 NA NA NA
12 22332 22417 -11154 22308 0.2307 1 0.631
16 22335 22447 -11151 22303 5.751 4 0.2185
26 22350 22533 -11149 22298 4.402 10 0.9274

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.721 0.1775 -20.97 -4.069 -3.374 fixed
poly(age, 3, raw = TRUE)1 0.5982 0.03643 16.42 0.5268 0.6696 fixed
poly(age, 3, raw = TRUE)2 -0.02748 0.00231 -11.9 -0.03201 -0.02296 fixed
poly(age, 3, raw = TRUE)3 0.000405 0.0000455 8.901 0.0003159 0.0004942 fixed
male -0.08998 0.01922 -4.682 -0.1276 -0.05231 fixed
sibling_count3 -0.0135 0.0288 -0.4689 -0.06995 0.04294 fixed
sibling_count4 -0.07019 0.03389 -2.071 -0.1366 -0.003757 fixed
sibling_count5 -0.08416 0.03975 -2.118 -0.1621 -0.006265 fixed
sibling_count5+ -0.2069 0.03541 -5.841 -0.2763 -0.1374 fixed
sd_(Intercept).mother_pidlink 0.5042 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7715 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.669 0.1792 -20.47 -4.02 -3.318 fixed
birth_order -0.01542 0.007277 -2.119 -0.02968 -0.001158 fixed
poly(age, 3, raw = TRUE)1 0.5933 0.03651 16.25 0.5217 0.6648 fixed
poly(age, 3, raw = TRUE)2 -0.02725 0.002312 -11.79 -0.03179 -0.02272 fixed
poly(age, 3, raw = TRUE)3 0.000401 0.00004554 8.805 0.0003118 0.0004903 fixed
male -0.0895 0.01922 -4.657 -0.1272 -0.05183 fixed
sibling_count3 -0.003641 0.02916 -0.1249 -0.06079 0.05351 fixed
sibling_count4 -0.04804 0.03545 -1.355 -0.1175 0.02143 fixed
sibling_count5 -0.04937 0.04297 -1.149 -0.1336 0.03486 fixed
sibling_count5+ -0.1363 0.0486 -2.804 -0.2316 -0.04105 fixed
sd_(Intercept).mother_pidlink 0.5031 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7718 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.707 0.1791 -20.7 -4.058 -3.356 fixed
poly(age, 3, raw = TRUE)1 0.5959 0.03652 16.32 0.5243 0.6675 fixed
poly(age, 3, raw = TRUE)2 -0.02739 0.002312 -11.85 -0.03192 -0.02286 fixed
poly(age, 3, raw = TRUE)3 0.0004033 0.00004554 8.857 0.000314 0.0004925 fixed
male -0.09039 0.01922 -4.703 -0.1281 -0.05272 fixed
sibling_count3 -0.01831 0.03031 -0.6043 -0.07772 0.04109 fixed
sibling_count4 -0.06181 0.03823 -1.617 -0.1367 0.01312 fixed
sibling_count5 -0.03862 0.04684 -0.8245 -0.1304 0.05318 fixed
sibling_count5+ -0.1467 0.05156 -2.846 -0.2478 -0.04566 fixed
birth_order_nonlinear2 0.005876 0.02377 0.2472 -0.04072 0.05247 fixed
birth_order_nonlinear3 0.01979 0.03171 0.624 -0.04237 0.08195 fixed
birth_order_nonlinear4 -0.04034 0.04163 -0.969 -0.1219 0.04125 fixed
birth_order_nonlinear5 -0.1249 0.05231 -2.388 -0.2275 -0.02241 fixed
birth_order_nonlinear5+ -0.06325 0.05686 -1.112 -0.1747 0.04819 fixed
sd_(Intercept).mother_pidlink 0.5046 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7711 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.714 0.1801 -20.62 -4.067 -3.361 fixed
poly(age, 3, raw = TRUE)1 0.5966 0.03659 16.31 0.5249 0.6684 fixed
poly(age, 3, raw = TRUE)2 -0.02745 0.002316 -11.85 -0.03199 -0.02291 fixed
poly(age, 3, raw = TRUE)3 0.0004048 0.00004562 8.874 0.0003154 0.0004942 fixed
male -0.08969 0.01923 -4.664 -0.1274 -0.05199 fixed
count_birth_order2/2 0.01737 0.03542 0.4903 -0.05206 0.0868 fixed
count_birth_order1/3 -0.0044 0.03957 -0.1112 -0.08195 0.07316 fixed
count_birth_order2/3 -0.01004 0.0387 -0.2595 -0.08589 0.06581 fixed
count_birth_order3/3 -0.004255 0.04242 -0.1003 -0.08739 0.07888 fixed
count_birth_order1/4 -0.1121 0.05889 -1.903 -0.2275 0.003338 fixed
count_birth_order2/4 -0.06207 0.05288 -1.174 -0.1657 0.04158 fixed
count_birth_order3/4 -0.05365 0.05062 -1.06 -0.1529 0.04557 fixed
count_birth_order4/4 -0.03904 0.05121 -0.7624 -0.1394 0.06132 fixed
count_birth_order1/5 0.06733 0.08556 0.787 -0.1004 0.235 fixed
count_birth_order2/5 -0.03288 0.07661 -0.4292 -0.183 0.1173 fixed
count_birth_order3/5 0.02536 0.06882 0.3685 -0.1095 0.1602 fixed
count_birth_order4/5 -0.1427 0.0634 -2.251 -0.267 -0.01844 fixed
count_birth_order5/5 -0.167 0.0611 -2.733 -0.2867 -0.04723 fixed
count_birth_order1/5+ -0.1296 0.09156 -1.416 -0.3091 0.04982 fixed
count_birth_order2/5+ -0.154 0.1017 -1.514 -0.3532 0.04533 fixed
count_birth_order3/5+ -0.08844 0.07879 -1.122 -0.2429 0.06599 fixed
count_birth_order4/5+ -0.2195 0.07233 -3.034 -0.3612 -0.07771 fixed
count_birth_order5/5+ -0.2618 0.063 -4.156 -0.3853 -0.1383 fixed
count_birth_order5+/5+ -0.2058 0.04418 -4.659 -0.2924 -0.1192 fixed
sd_(Intercept).mother_pidlink 0.5045 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7713 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 22622 22699 -11300 22600 NA NA NA
12 22619 22704 -11298 22595 4.494 1 0.03401
16 22623 22736 -11296 22591 4.198 4 0.3799
26 22635 22819 -11292 22583 7.659 10 0.6621

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.693 0.1793 -20.6 -4.044 -3.341 fixed
poly(age, 3, raw = TRUE)1 0.5935 0.03678 16.14 0.5214 0.6656 fixed
poly(age, 3, raw = TRUE)2 -0.02715 0.002329 -11.65 -0.03172 -0.02258 fixed
poly(age, 3, raw = TRUE)3 0.0003988 0.00004585 8.698 0.0003089 0.0004887 fixed
male -0.08566 0.01947 -4.399 -0.1238 -0.04749 fixed
sibling_count3 -0.03641 0.02752 -1.323 -0.09036 0.01753 fixed
sibling_count4 -0.09427 0.03407 -2.767 -0.1611 -0.02749 fixed
sibling_count5 -0.2168 0.04512 -4.805 -0.3052 -0.1284 fixed
sibling_count5+ -0.2744 0.04164 -6.59 -0.356 -0.1928 fixed
sd_(Intercept).mother_pidlink 0.5036 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7663 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.688 0.1818 -20.29 -4.044 -3.332 fixed
birth_order -0.001416 0.008884 -0.1594 -0.01883 0.016 fixed
poly(age, 3, raw = TRUE)1 0.5931 0.0369 16.07 0.5208 0.6654 fixed
poly(age, 3, raw = TRUE)2 -0.02713 0.002334 -11.62 -0.0317 -0.02255 fixed
poly(age, 3, raw = TRUE)3 0.0003984 0.00004593 8.674 0.0003084 0.0004884 fixed
male -0.08565 0.01948 -4.398 -0.1238 -0.04747 fixed
sibling_count3 -0.03547 0.02816 -1.26 -0.09066 0.01972 fixed
sibling_count4 -0.09217 0.03655 -2.522 -0.1638 -0.02053 fixed
sibling_count5 -0.2133 0.05003 -4.264 -0.3114 -0.1153 fixed
sibling_count5+ -0.2677 0.05947 -4.501 -0.3842 -0.1511 fixed
sd_(Intercept).mother_pidlink 0.5036 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7664 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.713 0.1813 -20.49 -4.068 -3.358 fixed
poly(age, 3, raw = TRUE)1 0.5961 0.03692 16.15 0.5237 0.6684 fixed
poly(age, 3, raw = TRUE)2 -0.0273 0.002334 -11.69 -0.03187 -0.02272 fixed
poly(age, 3, raw = TRUE)3 0.0004017 0.00004594 8.745 0.0003117 0.0004918 fixed
male -0.0861 0.01948 -4.42 -0.1243 -0.04792 fixed
sibling_count3 -0.03912 0.02934 -1.333 -0.09663 0.01839 fixed
sibling_count4 -0.1015 0.03929 -2.584 -0.1785 -0.02451 fixed
sibling_count5 -0.1942 0.05459 -3.558 -0.3012 -0.08722 fixed
sibling_count5+ -0.2932 0.0642 -4.567 -0.419 -0.1674 fixed
birth_order_nonlinear2 0.01886 0.02302 0.819 -0.02627 0.06398 fixed
birth_order_nonlinear3 0.0077 0.03246 0.2372 -0.05593 0.07133 fixed
birth_order_nonlinear4 0.02221 0.04513 0.4921 -0.06624 0.1107 fixed
birth_order_nonlinear5 -0.07965 0.06213 -1.282 -0.2014 0.04212 fixed
birth_order_nonlinear5+ 0.0613 0.07122 0.8607 -0.0783 0.2009 fixed
sd_(Intercept).mother_pidlink 0.5036 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7663 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.719 0.1822 -20.41 -4.076 -3.361 fixed
poly(age, 3, raw = TRUE)1 0.5962 0.03702 16.11 0.5237 0.6688 fixed
poly(age, 3, raw = TRUE)2 -0.02733 0.002341 -11.67 -0.03191 -0.02274 fixed
poly(age, 3, raw = TRUE)3 0.0004025 0.00004607 8.737 0.0003122 0.0004928 fixed
male -0.0853 0.01949 -4.376 -0.1235 -0.04709 fixed
count_birth_order2/2 0.03444 0.03202 1.076 -0.02832 0.0972 fixed
count_birth_order1/3 -0.02468 0.03831 -0.6441 -0.09977 0.05041 fixed
count_birth_order2/3 -0.02179 0.0368 -0.5921 -0.09391 0.05033 fixed
count_birth_order3/3 -0.02656 0.04033 -0.6585 -0.1056 0.05249 fixed
count_birth_order1/4 -0.08614 0.06046 -1.425 -0.2046 0.03235 fixed
count_birth_order2/4 -0.09269 0.05345 -1.734 -0.1974 0.01207 fixed
count_birth_order3/4 -0.1176 0.05075 -2.318 -0.2171 -0.01819 fixed
count_birth_order4/4 -0.03975 0.0506 -0.7855 -0.1389 0.05943 fixed
count_birth_order1/5 -0.1523 0.1045 -1.458 -0.3571 0.05249 fixed
count_birth_order2/5 -0.1503 0.09998 -1.503 -0.3462 0.04566 fixed
count_birth_order3/5 -0.1079 0.08106 -1.331 -0.2668 0.05096 fixed
count_birth_order4/5 -0.2259 0.06879 -3.285 -0.3608 -0.09113 fixed
count_birth_order5/5 -0.2732 0.06725 -4.063 -0.4051 -0.1414 fixed
count_birth_order1/5+ -0.3269 0.1345 -2.43 -0.5906 -0.06329 fixed
count_birth_order2/5+ -0.2724 0.1242 -2.194 -0.5157 -0.02903 fixed
count_birth_order3/5+ -0.2516 0.1016 -2.476 -0.4508 -0.05242 fixed
count_birth_order4/5+ -0.278 0.09641 -2.883 -0.467 -0.08903 fixed
count_birth_order5/5+ -0.3623 0.07667 -4.725 -0.5126 -0.212 fixed
count_birth_order5+/5+ -0.2261 0.05061 -4.468 -0.3253 -0.1269 fixed
sd_(Intercept).mother_pidlink 0.5028 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.767 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 21671 21748 -10824 21649 NA NA NA
12 21673 21757 -10824 21649 0.02562 1 0.8728
16 21675 21787 -10822 21643 5.499 4 0.2398
26 21691 21873 -10820 21639 4.027 10 0.9461

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

g-factor 2007 old

birthorder <- birthorder %>% mutate(outcome = g_factor_2007_old)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.967 1.932 -3.605 -10.75 -3.179 fixed
poly(age, 3, raw = TRUE)1 0.7971 0.2008 3.969 0.4035 1.191 fixed
poly(age, 3, raw = TRUE)2 -0.02803 0.006866 -4.083 -0.04149 -0.01458 fixed
poly(age, 3, raw = TRUE)3 0.0003121 0.00007722 4.042 0.0001608 0.0004635 fixed
male 0.03153 0.02044 1.542 -0.008534 0.07159 fixed
sibling_count3 0.007267 0.05738 0.1266 -0.1052 0.1197 fixed
sibling_count4 -0.1355 0.0562 -2.412 -0.2457 -0.02541 fixed
sibling_count5 -0.04119 0.0572 -0.7201 -0.1533 0.07092 fixed
sibling_count5+ -0.2903 0.04694 -6.184 -0.3823 -0.1983 fixed
sd_(Intercept).mother_pidlink 0.5938 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7824 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.953 1.933 -3.597 -10.74 -3.164 fixed
birth_order -0.001446 0.004365 -0.3312 -0.01 0.00711 fixed
poly(age, 3, raw = TRUE)1 0.7962 0.2009 3.964 0.4025 1.19 fixed
poly(age, 3, raw = TRUE)2 -0.02801 0.006867 -4.078 -0.04147 -0.01455 fixed
poly(age, 3, raw = TRUE)3 0.0003118 0.00007724 4.037 0.0001604 0.0004632 fixed
male 0.03162 0.02044 1.547 -0.008452 0.07169 fixed
sibling_count3 0.007867 0.05741 0.137 -0.1047 0.1204 fixed
sibling_count4 -0.1343 0.05633 -2.384 -0.2447 -0.02388 fixed
sibling_count5 -0.03904 0.05757 -0.6781 -0.1519 0.0738 fixed
sibling_count5+ -0.2846 0.04998 -5.695 -0.3826 -0.1867 fixed
sd_(Intercept).mother_pidlink 0.5937 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7825 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.89 1.934 -3.562 -10.68 -3.099 fixed
poly(age, 3, raw = TRUE)1 0.7923 0.201 3.942 0.3984 1.186 fixed
poly(age, 3, raw = TRUE)2 -0.02789 0.006872 -4.059 -0.04136 -0.01442 fixed
poly(age, 3, raw = TRUE)3 0.0003108 0.00007729 4.021 0.0001593 0.0004623 fixed
male 0.03107 0.02045 1.52 -0.009006 0.07115 fixed
sibling_count3 0.01322 0.05816 0.2273 -0.1008 0.1272 fixed
sibling_count4 -0.1258 0.05741 -2.192 -0.2384 -0.01333 fixed
sibling_count5 -0.02862 0.05909 -0.4843 -0.1444 0.08719 fixed
sibling_count5+ -0.2855 0.05142 -5.551 -0.3862 -0.1847 fixed
birth_order_nonlinear2 -0.062 0.03227 -1.921 -0.1252 0.001245 fixed
birth_order_nonlinear3 -0.05099 0.0347 -1.469 -0.119 0.01702 fixed
birth_order_nonlinear4 -0.04717 0.03805 -1.24 -0.1217 0.0274 fixed
birth_order_nonlinear5 -0.04488 0.04311 -1.041 -0.1294 0.03961 fixed
birth_order_nonlinear5+ -0.02096 0.03744 -0.5599 -0.09435 0.05242 fixed
sd_(Intercept).mother_pidlink 0.5941 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7822 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.963 1.936 -3.596 -10.76 -3.168 fixed
poly(age, 3, raw = TRUE)1 0.8025 0.2012 3.989 0.4082 1.197 fixed
poly(age, 3, raw = TRUE)2 -0.02824 0.006877 -4.107 -0.04172 -0.01476 fixed
poly(age, 3, raw = TRUE)3 0.0003149 0.00007735 4.072 0.0001633 0.0004665 fixed
male 0.03059 0.02047 1.495 -0.009522 0.0707 fixed
count_birth_order2/2 -0.1199 0.07505 -1.598 -0.267 0.02718 fixed
count_birth_order1/3 -0.01534 0.07666 -0.2002 -0.1656 0.1349 fixed
count_birth_order2/3 -0.01554 0.08162 -0.1904 -0.1755 0.1444 fixed
count_birth_order3/3 -0.1157 0.08194 -1.412 -0.2763 0.04492 fixed
count_birth_order1/4 -0.1356 0.07888 -1.72 -0.2902 0.01895 fixed
count_birth_order2/4 -0.1728 0.08201 -2.107 -0.3335 -0.01202 fixed
count_birth_order3/4 -0.3025 0.08428 -3.589 -0.4677 -0.1373 fixed
count_birth_order4/4 -0.1715 0.08514 -2.015 -0.3384 -0.004665 fixed
count_birth_order1/5 -0.05674 0.08618 -0.6584 -0.2256 0.1122 fixed
count_birth_order2/5 -0.1972 0.08763 -2.25 -0.3689 -0.0254 fixed
count_birth_order3/5 -0.05314 0.08644 -0.6147 -0.2226 0.1163 fixed
count_birth_order4/5 -0.1348 0.08841 -1.525 -0.3081 0.03846 fixed
count_birth_order5/5 -0.02949 0.09025 -0.3268 -0.2064 0.1474 fixed
count_birth_order1/5+ -0.3319 0.06889 -4.818 -0.4669 -0.1969 fixed
count_birth_order2/5+ -0.3771 0.06928 -5.443 -0.5129 -0.2413 fixed
count_birth_order3/5+ -0.3177 0.06806 -4.669 -0.4511 -0.1843 fixed
count_birth_order4/5+ -0.359 0.06685 -5.371 -0.49 -0.228 fixed
count_birth_order5/5+ -0.376 0.06763 -5.559 -0.5085 -0.2434 fixed
count_birth_order5+/5+ -0.3317 0.05973 -5.553 -0.4487 -0.2146 fixed
sd_(Intercept).mother_pidlink 0.5925 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7828 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 21826 21903 -10902 21804 NA NA NA
12 21828 21912 -10902 21804 0.1104 1 0.7396
16 21831 21943 -10900 21799 4.713 4 0.3181
26 21840 22021 -10894 21788 11.7 10 0.3055

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.205 3.657 -1.697 -13.37 0.9634 fixed
poly(age, 3, raw = TRUE)1 0.7234 0.3805 1.901 -0.02235 1.469 fixed
poly(age, 3, raw = TRUE)2 -0.02532 0.01301 -1.947 -0.05081 0.0001718 fixed
poly(age, 3, raw = TRUE)3 0.0002839 0.0001461 1.943 -0.000002515 0.0005703 fixed
male -0.03673 0.02918 -1.259 -0.09393 0.02046 fixed
sibling_count3 0.01766 0.05501 0.3211 -0.09015 0.1255 fixed
sibling_count4 -0.08957 0.05691 -1.574 -0.2011 0.02198 fixed
sibling_count5 -0.1664 0.06228 -2.671 -0.2885 -0.04431 fixed
sibling_count5+ -0.3372 0.0539 -6.256 -0.4429 -0.2316 fixed
sd_(Intercept).mother_pidlink 0.539 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7449 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.448 3.656 -1.764 -13.61 0.7171 fixed
birth_order 0.023 0.009659 2.381 0.004065 0.04193 fixed
poly(age, 3, raw = TRUE)1 0.7442 0.3803 1.957 -0.001197 1.49 fixed
poly(age, 3, raw = TRUE)2 -0.02604 0.013 -2.003 -0.05152 -0.0005649 fixed
poly(age, 3, raw = TRUE)3 0.000293 0.000146 2.006 0.000006727 0.0005792 fixed
male -0.03895 0.02917 -1.335 -0.09613 0.01823 fixed
sibling_count3 0.007197 0.05516 0.1305 -0.1009 0.1153 fixed
sibling_count4 -0.1165 0.05801 -2.009 -0.2302 -0.002827 fixed
sibling_count5 -0.2102 0.06492 -3.237 -0.3374 -0.08293 fixed
sibling_count5+ -0.4215 0.06445 -6.539 -0.5478 -0.2951 fixed
sd_(Intercept).mother_pidlink 0.5393 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7441 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.529 3.662 -1.783 -13.71 0.648 fixed
poly(age, 3, raw = TRUE)1 0.7555 0.381 1.983 0.008815 1.502 fixed
poly(age, 3, raw = TRUE)2 -0.02643 0.01302 -2.03 -0.05195 -0.0009143 fixed
poly(age, 3, raw = TRUE)3 0.0002976 0.0001463 2.034 0.00001085 0.0005843 fixed
male -0.03853 0.02917 -1.321 -0.09571 0.01864 fixed
sibling_count3 0.002132 0.05581 0.03821 -0.1073 0.1115 fixed
sibling_count4 -0.124 0.0598 -2.074 -0.2412 -0.006812 fixed
sibling_count5 -0.2096 0.06759 -3.101 -0.342 -0.07709 fixed
sibling_count5+ -0.4414 0.06577 -6.711 -0.5703 -0.3125 fixed
birth_order_nonlinear2 0.009153 0.03802 0.2407 -0.06536 0.08367 fixed
birth_order_nonlinear3 0.06512 0.04521 1.441 -0.02348 0.1537 fixed
birth_order_nonlinear4 0.07384 0.05563 1.327 -0.03519 0.1829 fixed
birth_order_nonlinear5 0.0474 0.06735 0.7038 -0.08461 0.1794 fixed
birth_order_nonlinear5+ 0.1984 0.06874 2.886 0.06366 0.3331 fixed
sd_(Intercept).mother_pidlink 0.5393 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7441 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.77 3.663 -1.848 -13.95 0.4091 fixed
poly(age, 3, raw = TRUE)1 0.7853 0.381 2.061 0.03849 1.532 fixed
poly(age, 3, raw = TRUE)2 -0.02747 0.01302 -2.11 -0.05299 -0.001948 fixed
poly(age, 3, raw = TRUE)3 0.0003095 0.0001463 2.115 0.0000227 0.0005962 fixed
male -0.03938 0.02918 -1.349 -0.09658 0.01782 fixed
count_birth_order2/2 -0.1061 0.07239 -1.466 -0.248 0.03578 fixed
count_birth_order1/3 -0.04743 0.06861 -0.6912 -0.1819 0.08705 fixed
count_birth_order2/3 0.0001341 0.07494 0.001789 -0.1467 0.147 fixed
count_birth_order3/3 0.002427 0.08289 0.02927 -0.16 0.1649 fixed
count_birth_order1/4 -0.1624 0.07978 -2.036 -0.3188 -0.006039 fixed
count_birth_order2/4 -0.1517 0.08132 -1.865 -0.3111 0.007689 fixed
count_birth_order3/4 -0.203 0.08231 -2.466 -0.3643 -0.04165 fixed
count_birth_order4/4 0.02222 0.08949 0.2483 -0.1532 0.1976 fixed
count_birth_order1/5 -0.258 0.1027 -2.511 -0.4593 -0.05663 fixed
count_birth_order2/5 -0.3278 0.1062 -3.088 -0.5359 -0.1197 fixed
count_birth_order3/5 -0.05324 0.09669 -0.5506 -0.2427 0.1363 fixed
count_birth_order4/5 -0.2332 0.09627 -2.423 -0.4219 -0.04454 fixed
count_birth_order5/5 -0.2 0.1008 -1.984 -0.3975 -0.002421 fixed
count_birth_order1/5+ -0.5937 0.0977 -6.077 -0.7852 -0.4022 fixed
count_birth_order2/5+ -0.361 0.09474 -3.811 -0.5467 -0.1753 fixed
count_birth_order3/5+ -0.3577 0.0892 -4.01 -0.5326 -0.1829 fixed
count_birth_order4/5+ -0.471 0.08863 -5.314 -0.6447 -0.2973 fixed
count_birth_order5/5+ -0.4359 0.08616 -5.059 -0.6048 -0.267 fixed
count_birth_order5+/5+ -0.2855 0.06908 -4.133 -0.4209 -0.1501 fixed
sd_(Intercept).mother_pidlink 0.5405 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7422 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 9408 9476 -4693 9386 NA NA NA
12 9404 9478 -4690 9380 5.676 1 0.01719
16 9408 9507 -4688 9376 4.108 4 0.3915
26 9409 9570 -4679 9357 18.76 10 0.04348

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.46 3.656 -1.767 -13.63 0.7058 fixed
poly(age, 3, raw = TRUE)1 0.7521 0.3804 1.977 0.006589 1.498 fixed
poly(age, 3, raw = TRUE)2 -0.02635 0.013 -2.026 -0.05183 -0.0008642 fixed
poly(age, 3, raw = TRUE)3 0.0002955 0.0001461 2.023 0.000009182 0.0005818 fixed
male -0.03596 0.02916 -1.233 -0.09311 0.02119 fixed
sibling_count3 -0.01256 0.06097 -0.2061 -0.1321 0.1069 fixed
sibling_count4 -0.06261 0.06183 -1.013 -0.1838 0.05857 fixed
sibling_count5 -0.1069 0.06451 -1.657 -0.2333 0.01954 fixed
sibling_count5+ -0.2746 0.0561 -4.896 -0.3846 -0.1647 fixed
sd_(Intercept).mother_pidlink 0.5448 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7446 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.577 3.657 -1.799 -13.74 0.5897 fixed
birth_order 0.009836 0.008549 1.151 -0.006918 0.02659 fixed
poly(age, 3, raw = TRUE)1 0.7626 0.3804 2.005 0.01701 1.508 fixed
poly(age, 3, raw = TRUE)2 -0.02672 0.013 -2.055 -0.0522 -0.001234 fixed
poly(age, 3, raw = TRUE)3 0.0003002 0.0001461 2.055 0.00001383 0.0005865 fixed
male -0.03681 0.02917 -1.262 -0.09397 0.02035 fixed
sibling_count3 -0.01729 0.06112 -0.2829 -0.1371 0.1025 fixed
sibling_count4 -0.07328 0.06253 -1.172 -0.1958 0.04928 fixed
sibling_count5 -0.1237 0.06615 -1.87 -0.2533 0.005958 fixed
sibling_count5+ -0.31 0.06394 -4.847 -0.4353 -0.1846 fixed
sd_(Intercept).mother_pidlink 0.5453 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7443 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.802 3.663 -1.857 -13.98 0.3766 fixed
poly(age, 3, raw = TRUE)1 0.787 0.381 2.065 0.04012 1.534 fixed
poly(age, 3, raw = TRUE)2 -0.02754 0.01302 -2.115 -0.05307 -0.002019 fixed
poly(age, 3, raw = TRUE)3 0.0003094 0.0001463 2.115 0.00002266 0.0005962 fixed
male -0.03809 0.02917 -1.306 -0.09527 0.01909 fixed
sibling_count3 -0.01862 0.06182 -0.3012 -0.1398 0.1026 fixed
sibling_count4 -0.07304 0.06402 -1.141 -0.1985 0.05243 fixed
sibling_count5 -0.1131 0.06838 -1.654 -0.2472 0.0209 fixed
sibling_count5+ -0.3226 0.06534 -4.937 -0.4507 -0.1945 fixed
birth_order_nonlinear2 0.006261 0.03896 0.1607 -0.0701 0.08262 fixed
birth_order_nonlinear3 0.02381 0.04523 0.5263 -0.06485 0.1125 fixed
birth_order_nonlinear4 0.01933 0.05432 0.3559 -0.08713 0.1258 fixed
birth_order_nonlinear5 -0.01933 0.06473 -0.2986 -0.1462 0.1075 fixed
birth_order_nonlinear5+ 0.1189 0.06293 1.89 -0.004426 0.2422 fixed
sd_(Intercept).mother_pidlink 0.5463 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7437 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -7.278 3.674 -1.981 -14.48 -0.07648 fixed
poly(age, 3, raw = TRUE)1 0.8385 0.3822 2.194 0.08932 1.588 fixed
poly(age, 3, raw = TRUE)2 -0.02931 0.01306 -2.244 -0.05491 -0.003709 fixed
poly(age, 3, raw = TRUE)3 0.0003294 0.0001467 2.245 0.00004183 0.000617 fixed
male -0.03995 0.02924 -1.366 -0.09725 0.01735 fixed
count_birth_order2/2 -0.04316 0.08217 -0.5253 -0.2042 0.1179 fixed
count_birth_order1/3 -0.06737 0.07659 -0.8796 -0.2175 0.08275 fixed
count_birth_order2/3 0.01362 0.08304 0.164 -0.1491 0.1764 fixed
count_birth_order3/3 -0.008673 0.09006 -0.0963 -0.1852 0.1678 fixed
count_birth_order1/4 -0.06541 0.08479 -0.7714 -0.2316 0.1008 fixed
count_birth_order2/4 -0.08562 0.08554 -1.001 -0.2533 0.08204 fixed
count_birth_order3/4 -0.1709 0.09005 -1.898 -0.3474 0.005567 fixed
count_birth_order4/4 0.01303 0.09927 0.1312 -0.1815 0.2076 fixed
count_birth_order1/5 -0.1005 0.09729 -1.033 -0.2912 0.0902 fixed
count_birth_order2/5 -0.1612 0.1008 -1.599 -0.3589 0.03638 fixed
count_birth_order3/5 -0.09824 0.1004 -0.9789 -0.2949 0.09845 fixed
count_birth_order4/5 -0.2109 0.101 -2.089 -0.4088 -0.01305 fixed
count_birth_order5/5 -0.05866 0.1047 -0.5602 -0.2639 0.1466 fixed
count_birth_order1/5+ -0.3759 0.08989 -4.182 -0.5521 -0.1997 fixed
count_birth_order2/5+ -0.3399 0.09127 -3.724 -0.5187 -0.161 fixed
count_birth_order3/5+ -0.2308 0.08689 -2.657 -0.4011 -0.06054 fixed
count_birth_order4/5+ -0.3191 0.086 -3.71 -0.4876 -0.1505 fixed
count_birth_order5/5+ -0.4139 0.08772 -4.718 -0.5858 -0.2419 fixed
count_birth_order5+/5+ -0.2232 0.0714 -3.126 -0.3631 -0.08326 fixed
sd_(Intercept).mother_pidlink 0.5457 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7438 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 9482 9550 -4730 9460 NA NA NA
12 9483 9557 -4729 9459 1.325 1 0.2496
16 9487 9586 -4728 9455 3.804 4 0.4332
26 9496 9657 -4722 9444 11.35 10 0.3306

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -7.076 3.679 -1.923 -14.29 0.1352 fixed
poly(age, 3, raw = TRUE)1 0.8127 0.3827 2.124 0.06267 1.563 fixed
poly(age, 3, raw = TRUE)2 -0.02836 0.01307 -2.169 -0.05398 -0.002738 fixed
poly(age, 3, raw = TRUE)3 0.0003175 0.0001468 2.162 0.00002971 0.0006053 fixed
male -0.04285 0.02939 -1.458 -0.1004 0.01475 fixed
sibling_count3 0.02927 0.05419 0.5401 -0.07694 0.1355 fixed
sibling_count4 -0.04721 0.05629 -0.8387 -0.1575 0.06311 fixed
sibling_count5 -0.1309 0.06315 -2.073 -0.2547 -0.007122 fixed
sibling_count5+ -0.2998 0.05394 -5.557 -0.4055 -0.1941 fixed
sd_(Intercept).mother_pidlink 0.5428 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7428 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -7.32 3.679 -1.99 -14.53 -0.1093 fixed
birth_order 0.02161 0.009902 2.182 0.002199 0.04101 fixed
poly(age, 3, raw = TRUE)1 0.8336 0.3825 2.179 0.08381 1.583 fixed
poly(age, 3, raw = TRUE)2 -0.02909 0.01307 -2.226 -0.0547 -0.003472 fixed
poly(age, 3, raw = TRUE)3 0.0003266 0.0001468 2.225 0.00003884 0.0006143 fixed
male -0.04449 0.02938 -1.514 -0.1021 0.01309 fixed
sibling_count3 0.01921 0.05436 0.3534 -0.08733 0.1258 fixed
sibling_count4 -0.07241 0.05744 -1.261 -0.185 0.04017 fixed
sibling_count5 -0.1701 0.06564 -2.592 -0.2988 -0.0415 fixed
sibling_count5+ -0.3787 0.06491 -5.833 -0.5059 -0.2514 fixed
sd_(Intercept).mother_pidlink 0.5429 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7423 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -7.42 3.683 -2.015 -14.64 -0.202 fixed
poly(age, 3, raw = TRUE)1 0.8467 0.383 2.211 0.09602 1.597 fixed
poly(age, 3, raw = TRUE)2 -0.02955 0.01308 -2.258 -0.05519 -0.003901 fixed
poly(age, 3, raw = TRUE)3 0.000332 0.000147 2.259 0.00004392 0.00062 fixed
male -0.04381 0.02937 -1.492 -0.1014 0.01376 fixed
sibling_count3 0.01676 0.05507 0.3044 -0.09117 0.1247 fixed
sibling_count4 -0.07309 0.05928 -1.233 -0.1893 0.0431 fixed
sibling_count5 -0.1642 0.06813 -2.41 -0.2978 -0.03069 fixed
sibling_count5+ -0.4014 0.06633 -6.051 -0.5314 -0.2714 fixed
birth_order_nonlinear2 0.01263 0.03785 0.3338 -0.06155 0.08682 fixed
birth_order_nonlinear3 0.0498 0.045 1.107 -0.03841 0.138 fixed
birth_order_nonlinear4 0.05286 0.05643 0.9368 -0.05774 0.1635 fixed
birth_order_nonlinear5 0.0456 0.0695 0.6561 -0.09062 0.1818 fixed
birth_order_nonlinear5+ 0.2082 0.07078 2.941 0.06945 0.3469 fixed
sd_(Intercept).mother_pidlink 0.5432 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.742 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -7.49 3.684 -2.033 -14.71 -0.2699 fixed
poly(age, 3, raw = TRUE)1 0.8596 0.3831 2.244 0.1088 1.61 fixed
poly(age, 3, raw = TRUE)2 -0.03005 0.01309 -2.296 -0.0557 -0.004402 fixed
poly(age, 3, raw = TRUE)3 0.0003384 0.000147 2.302 0.00005028 0.0006264 fixed
male -0.04343 0.0294 -1.477 -0.1011 0.0142 fixed
count_birth_order2/2 -0.08876 0.07016 -1.265 -0.2263 0.04875 fixed
count_birth_order1/3 -0.03279 0.06767 -0.4846 -0.1654 0.09984 fixed
count_birth_order2/3 0.03804 0.07462 0.5098 -0.1082 0.1843 fixed
count_birth_order3/3 -0.002867 0.08072 -0.03551 -0.1611 0.1553 fixed
count_birth_order1/4 -0.09423 0.07942 -1.186 -0.2499 0.06143 fixed
count_birth_order2/4 -0.09867 0.0809 -1.22 -0.2572 0.05988 fixed
count_birth_order3/4 -0.1327 0.08217 -1.614 -0.2937 0.0284 fixed
count_birth_order4/4 0.003595 0.0882 0.04076 -0.1693 0.1765 fixed
count_birth_order1/5 -0.2386 0.1019 -2.341 -0.4383 -0.03884 fixed
count_birth_order2/5 -0.3413 0.1077 -3.169 -0.5524 -0.1302 fixed
count_birth_order3/5 -0.01051 0.09789 -0.1074 -0.2024 0.1814 fixed
count_birth_order4/5 -0.1728 0.09876 -1.75 -0.3664 0.02074 fixed
count_birth_order5/5 -0.1006 0.1067 -0.9431 -0.3098 0.1085 fixed
count_birth_order1/5+ -0.5258 0.09991 -5.263 -0.7216 -0.33 fixed
count_birth_order2/5+ -0.2842 0.09785 -2.904 -0.4759 -0.09238 fixed
count_birth_order3/5+ -0.3577 0.08934 -4.004 -0.5328 -0.1826 fixed
count_birth_order4/5+ -0.4301 0.09133 -4.709 -0.6091 -0.251 fixed
count_birth_order5/5+ -0.4212 0.08699 -4.841 -0.5917 -0.2507 fixed
count_birth_order5+/5+ -0.2325 0.06976 -3.333 -0.3692 -0.0958 fixed
sd_(Intercept).mother_pidlink 0.5445 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7403 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 9245 9313 -4612 9223 NA NA NA
12 9243 9317 -4609 9219 4.77 1 0.02896
16 9246 9345 -4607 9214 4.686 4 0.3211
26 9249 9409 -4598 9197 17.2 10 0.07015

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

g-factor 2007 young

birthorder <- birthorder %>% mutate(outcome = g_factor_2007_young)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -7.75 0.4555 -17.02 -8.643 -6.857 fixed
poly(age, 3, raw = TRUE)1 0.8809 0.0619 14.23 0.7596 1.002 fixed
poly(age, 3, raw = TRUE)2 -0.02996 0.002746 -10.91 -0.03534 -0.02458 fixed
poly(age, 3, raw = TRUE)3 0.0003113 0.00004013 7.759 0.0002327 0.00039 fixed
male -0.05593 0.01849 -3.025 -0.09217 -0.01969 fixed
sibling_count3 0.003237 0.04136 0.07825 -0.07783 0.0843 fixed
sibling_count4 -0.1427 0.04228 -3.375 -0.2255 -0.05982 fixed
sibling_count5 -0.09449 0.04566 -2.069 -0.184 -0.004994 fixed
sibling_count5+ -0.307 0.03653 -8.404 -0.3786 -0.2354 fixed
sd_(Intercept).mother_pidlink 0.5847 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7533 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -7.749 0.4555 -17.01 -8.642 -6.857 fixed
birth_order 0.0005934 0.004302 0.1379 -0.007839 0.009026 fixed
poly(age, 3, raw = TRUE)1 0.8806 0.06194 14.22 0.7592 1.002 fixed
poly(age, 3, raw = TRUE)2 -0.02995 0.002748 -10.9 -0.03533 -0.02456 fixed
poly(age, 3, raw = TRUE)3 0.0003112 0.00004015 7.751 0.0002325 0.0003899 fixed
male -0.05597 0.01849 -3.027 -0.09222 -0.01973 fixed
sibling_count3 0.002934 0.04142 0.07082 -0.07825 0.08412 fixed
sibling_count4 -0.1434 0.04256 -3.368 -0.2268 -0.05994 fixed
sibling_count5 -0.09562 0.0464 -2.061 -0.1866 -0.004681 fixed
sibling_count5+ -0.3102 0.04325 -7.173 -0.395 -0.2254 fixed
sd_(Intercept).mother_pidlink 0.5847 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7533 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -7.8 0.4563 -17.1 -8.694 -6.906 fixed
poly(age, 3, raw = TRUE)1 0.8848 0.06196 14.28 0.7634 1.006 fixed
poly(age, 3, raw = TRUE)2 -0.03011 0.002748 -10.96 -0.0355 -0.02473 fixed
poly(age, 3, raw = TRUE)3 0.0003133 0.00004015 7.803 0.0002346 0.0003919 fixed
male -0.05596 0.01849 -3.026 -0.0922 -0.01972 fixed
sibling_count3 0.001285 0.04195 0.03064 -0.08093 0.0835 fixed
sibling_count4 -0.149 0.04393 -3.391 -0.2351 -0.06288 fixed
sibling_count5 -0.08941 0.04865 -1.838 -0.1848 0.005933 fixed
sibling_count5+ -0.3059 0.04589 -6.664 -0.3958 -0.2159 fixed
birth_order_nonlinear2 0.06189 0.02732 2.265 0.00834 0.1154 fixed
birth_order_nonlinear3 0.006896 0.0318 0.2168 -0.05544 0.06923 fixed
birth_order_nonlinear4 0.03344 0.03735 0.8952 -0.03977 0.1066 fixed
birth_order_nonlinear5 -0.02553 0.04219 -0.6052 -0.1082 0.05715 fixed
birth_order_nonlinear5+ 0.0216 0.03841 0.5622 -0.05369 0.09689 fixed
sd_(Intercept).mother_pidlink 0.5844 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7533 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -7.852 0.4568 -17.19 -8.747 -6.957 fixed
poly(age, 3, raw = TRUE)1 0.8906 0.06197 14.37 0.7692 1.012 fixed
poly(age, 3, raw = TRUE)2 -0.03034 0.002748 -11.04 -0.03573 -0.02495 fixed
poly(age, 3, raw = TRUE)3 0.0003159 0.00004014 7.87 0.0002372 0.0003946 fixed
male -0.05667 0.0185 -3.064 -0.09292 -0.02042 fixed
count_birth_order2/2 0.08138 0.05571 1.461 -0.02781 0.1906 fixed
count_birth_order1/3 0.005201 0.05054 0.1029 -0.09385 0.1043 fixed
count_birth_order2/3 0.0915 0.05554 1.648 -0.01735 0.2004 fixed
count_birth_order3/3 -0.01376 0.06223 -0.2212 -0.1357 0.1082 fixed
count_birth_order1/4 -0.1705 0.05732 -2.974 -0.2828 -0.05815 fixed
count_birth_order2/4 -0.09368 0.05859 -1.599 -0.2085 0.02115 fixed
count_birth_order3/4 -0.1673 0.06116 -2.735 -0.2872 -0.04741 fixed
count_birth_order4/4 -0.006893 0.06672 -0.1033 -0.1377 0.1239 fixed
count_birth_order1/5 -0.1206 0.0694 -1.738 -0.2567 0.01539 fixed
count_birth_order2/5 -0.06432 0.07317 -0.879 -0.2077 0.07909 fixed
count_birth_order3/5 -0.09166 0.06978 -1.314 -0.2284 0.04511 fixed
count_birth_order4/5 -0.031 0.07323 -0.4233 -0.1745 0.1125 fixed
count_birth_order5/5 -0.03769 0.0714 -0.5278 -0.1776 0.1023 fixed
count_birth_order1/5+ -0.2167 0.06404 -3.384 -0.3422 -0.09122 fixed
count_birth_order2/5+ -0.2459 0.06271 -3.921 -0.3688 -0.123 fixed
count_birth_order3/5+ -0.2307 0.0588 -3.923 -0.3459 -0.1154 fixed
count_birth_order4/5+ -0.3284 0.0553 -5.94 -0.4368 -0.2201 fixed
count_birth_order5/5+ -0.3523 0.05331 -6.608 -0.4568 -0.2478 fixed
count_birth_order5+/5+ -0.2805 0.0418 -6.711 -0.3625 -0.1986 fixed
sd_(Intercept).mother_pidlink 0.583 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7537 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 24501 24580 -12240 24479 NA NA NA
12 24503 24589 -12240 24479 0.01897 1 0.8905
16 24503 24617 -12236 24471 7.893 4 0.09557
26 24508 24693 -12228 24456 15.5 10 0.1149

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -8.275 0.9556 -8.659 -10.15 -6.402 fixed
poly(age, 3, raw = TRUE)1 0.9443 0.1428 6.611 0.6643 1.224 fixed
poly(age, 3, raw = TRUE)2 -0.03243 0.006949 -4.667 -0.04605 -0.01881 fixed
poly(age, 3, raw = TRUE)3 0.0003491 0.0001103 3.167 0.000133 0.0005652 fixed
male -0.08542 0.02318 -3.686 -0.1308 -0.04 fixed
sibling_count3 0.01038 0.03598 0.2884 -0.06015 0.08091 fixed
sibling_count4 -0.09441 0.0408 -2.314 -0.1744 -0.01444 fixed
sibling_count5 -0.1658 0.04931 -3.363 -0.2625 -0.06919 fixed
sibling_count5+ -0.252 0.04407 -5.719 -0.3384 -0.1657 fixed
sd_(Intercept).mother_pidlink 0.54 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7094 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -8.376 0.9564 -8.758 -10.25 -6.502 fixed
birth_order 0.01868 0.008975 2.081 0.001089 0.03627 fixed
poly(age, 3, raw = TRUE)1 0.9548 0.1428 6.684 0.6748 1.235 fixed
poly(age, 3, raw = TRUE)2 -0.03293 0.006949 -4.739 -0.04655 -0.01931 fixed
poly(age, 3, raw = TRUE)3 0.0003577 0.0001103 3.244 0.0001416 0.0005739 fixed
male -0.08599 0.02317 -3.712 -0.1314 -0.04058 fixed
sibling_count3 0.001063 0.03626 0.02931 -0.07 0.07213 fixed
sibling_count4 -0.117 0.04221 -2.771 -0.1997 -0.03423 fixed
sibling_count5 -0.2044 0.05268 -3.881 -0.3077 -0.1012 fixed
sibling_count5+ -0.3352 0.05947 -5.636 -0.4517 -0.2186 fixed
sd_(Intercept).mother_pidlink 0.5404 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7088 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -8.405 0.9573 -8.78 -10.28 -6.529 fixed
poly(age, 3, raw = TRUE)1 0.9598 0.143 6.713 0.6796 1.24 fixed
poly(age, 3, raw = TRUE)2 -0.03314 0.006955 -4.765 -0.04677 -0.01951 fixed
poly(age, 3, raw = TRUE)3 0.0003606 0.0001104 3.267 0.0001443 0.0005769 fixed
male -0.08615 0.02318 -3.717 -0.1316 -0.04072 fixed
sibling_count3 -0.001466 0.03707 -0.03955 -0.07413 0.0712 fixed
sibling_count4 -0.1162 0.04424 -2.626 -0.2029 -0.02945 fixed
sibling_count5 -0.2014 0.056 -3.596 -0.3111 -0.09159 fixed
sibling_count5+ -0.3416 0.0622 -5.492 -0.4635 -0.2197 fixed
birth_order_nonlinear2 0.04855 0.02829 1.716 -0.006898 0.104 fixed
birth_order_nonlinear3 0.04855 0.03718 1.306 -0.02431 0.1214 fixed
birth_order_nonlinear4 0.04994 0.04837 1.032 -0.04488 0.1447 fixed
birth_order_nonlinear5 0.07905 0.0628 1.259 -0.04403 0.2021 fixed
birth_order_nonlinear5+ 0.1498 0.06863 2.182 0.01526 0.2843 fixed
sd_(Intercept).mother_pidlink 0.5404 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.709 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -8.377 0.9586 -8.739 -10.26 -6.498 fixed
poly(age, 3, raw = TRUE)1 0.9561 0.1432 6.677 0.6754 1.237 fixed
poly(age, 3, raw = TRUE)2 -0.03296 0.006966 -4.731 -0.04661 -0.01931 fixed
poly(age, 3, raw = TRUE)3 0.0003578 0.0001105 3.237 0.0001411 0.0005744 fixed
male -0.08551 0.02321 -3.685 -0.131 -0.04003 fixed
count_birth_order2/2 0.03458 0.04653 0.7431 -0.05662 0.1258 fixed
count_birth_order1/3 -0.01329 0.0451 -0.2946 -0.1017 0.0751 fixed
count_birth_order2/3 0.06255 0.04894 1.278 -0.03337 0.1585 fixed
count_birth_order3/3 0.03125 0.05458 0.5726 -0.07571 0.1382 fixed
count_birth_order1/4 -0.1189 0.06051 -1.964 -0.2375 -0.0002613 fixed
count_birth_order2/4 -0.1218 0.06038 -2.017 -0.2401 -0.003439 fixed
count_birth_order3/4 -0.0876 0.06289 -1.393 -0.2109 0.03566 fixed
count_birth_order4/4 -0.01094 0.06259 -0.1747 -0.1336 0.1117 fixed
count_birth_order1/5 -0.1497 0.09123 -1.641 -0.3285 0.02912 fixed
count_birth_order2/5 -0.1942 0.09599 -2.023 -0.3823 -0.006023 fixed
count_birth_order3/5 -0.2281 0.08465 -2.694 -0.394 -0.06215 fixed
count_birth_order4/5 -0.1943 0.07964 -2.44 -0.3504 -0.0382 fixed
count_birth_order5/5 -0.06289 0.0773 -0.8136 -0.2144 0.08861 fixed
count_birth_order1/5+ -0.4313 0.1114 -3.872 -0.6497 -0.213 fixed
count_birth_order2/5+ -0.151 0.1092 -1.382 -0.3651 0.06312 fixed
count_birth_order3/5+ -0.1233 0.0948 -1.3 -0.3091 0.06252 fixed
count_birth_order4/5+ -0.3679 0.08502 -4.327 -0.5345 -0.2012 fixed
count_birth_order5/5+ -0.3301 0.07643 -4.319 -0.4799 -0.1803 fixed
count_birth_order5+/5+ -0.2001 0.05454 -3.669 -0.307 -0.09319 fixed
sd_(Intercept).mother_pidlink 0.5386 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7095 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 13668 13740 -6823 13646 NA NA NA
12 13665 13744 -6821 13641 4.336 1 0.03732
16 13671 13776 -6820 13639 2.188 4 0.7013
26 13676 13847 -6812 13624 15.01 10 0.1318

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -8.245 0.9545 -8.638 -10.12 -6.374 fixed
poly(age, 3, raw = TRUE)1 0.9423 0.1426 6.607 0.6627 1.222 fixed
poly(age, 3, raw = TRUE)2 -0.03235 0.00694 -4.661 -0.04595 -0.01874 fixed
poly(age, 3, raw = TRUE)3 0.0003477 0.0001101 3.157 0.0001318 0.0005635 fixed
male -0.08633 0.02311 -3.735 -0.1316 -0.04103 fixed
sibling_count3 -0.01 0.03871 -0.2584 -0.08587 0.06587 fixed
sibling_count4 -0.07904 0.04223 -1.872 -0.1618 0.003723 fixed
sibling_count5 -0.1368 0.04772 -2.866 -0.2303 -0.04323 fixed
sibling_count5+ -0.2154 0.04124 -5.222 -0.2962 -0.1345 fixed
sd_(Intercept).mother_pidlink 0.5413 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7093 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -8.302 0.9557 -8.687 -10.18 -6.429 fixed
birth_order 0.008746 0.00768 1.139 -0.006307 0.0238 fixed
poly(age, 3, raw = TRUE)1 0.9486 0.1427 6.647 0.6689 1.228 fixed
poly(age, 3, raw = TRUE)2 -0.03264 0.006944 -4.701 -0.04625 -0.01903 fixed
poly(age, 3, raw = TRUE)3 0.0003526 0.0001102 3.2 0.0001366 0.0005686 fixed
male -0.08669 0.02311 -3.751 -0.132 -0.0414 fixed
sibling_count3 -0.01448 0.03891 -0.3721 -0.09074 0.06178 fixed
sibling_count4 -0.08899 0.04312 -2.064 -0.1735 -0.00447 fixed
sibling_count5 -0.1535 0.04993 -3.074 -0.2513 -0.05561 fixed
sibling_count5+ -0.2514 0.05196 -4.838 -0.3532 -0.1495 fixed
sd_(Intercept).mother_pidlink 0.5416 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7091 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -8.348 0.9563 -8.729 -10.22 -6.474 fixed
poly(age, 3, raw = TRUE)1 0.953 0.1428 6.673 0.6731 1.233 fixed
poly(age, 3, raw = TRUE)2 -0.0328 0.006948 -4.721 -0.04642 -0.01918 fixed
poly(age, 3, raw = TRUE)3 0.0003543 0.0001103 3.213 0.0001382 0.0005704 fixed
male -0.08662 0.02312 -3.746 -0.1319 -0.0413 fixed
sibling_count3 -0.0152 0.0397 -0.3828 -0.09301 0.06262 fixed
sibling_count4 -0.09031 0.04495 -2.009 -0.1784 -0.002209 fixed
sibling_count5 -0.1564 0.05279 -2.962 -0.2598 -0.05289 fixed
sibling_count5+ -0.2465 0.05381 -4.581 -0.352 -0.1411 fixed
birth_order_nonlinear2 0.0624 0.02904 2.148 0.005474 0.1193 fixed
birth_order_nonlinear3 0.01888 0.0368 0.513 -0.05325 0.09101 fixed
birth_order_nonlinear4 0.04156 0.04626 0.8983 -0.04911 0.1322 fixed
birth_order_nonlinear5 0.06251 0.05747 1.088 -0.05012 0.1751 fixed
birth_order_nonlinear5+ 0.06343 0.05866 1.081 -0.05154 0.1784 fixed
sd_(Intercept).mother_pidlink 0.541 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7094 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -8.352 0.9571 -8.727 -10.23 -6.476 fixed
poly(age, 3, raw = TRUE)1 0.9532 0.143 6.668 0.6731 1.233 fixed
poly(age, 3, raw = TRUE)2 -0.03285 0.006955 -4.723 -0.04648 -0.01922 fixed
poly(age, 3, raw = TRUE)3 0.0003559 0.0001104 3.225 0.0001396 0.0005723 fixed
male -0.08717 0.02314 -3.767 -0.1325 -0.04181 fixed
count_birth_order2/2 0.08017 0.05117 1.567 -0.02011 0.1805 fixed
count_birth_order1/3 -0.02997 0.0487 -0.6153 -0.1254 0.06549 fixed
count_birth_order2/3 0.06372 0.05225 1.22 -0.03868 0.1661 fixed
count_birth_order3/3 0.03225 0.05864 0.5499 -0.08269 0.1472 fixed
count_birth_order1/4 -0.1036 0.06091 -1.701 -0.223 0.01577 fixed
count_birth_order2/4 -0.06062 0.06142 -0.9869 -0.181 0.05976 fixed
count_birth_order3/4 -0.09628 0.06447 -1.493 -0.2226 0.03007 fixed
count_birth_order4/4 0.04985 0.0662 0.7531 -0.0799 0.1796 fixed
count_birth_order1/5 -0.06508 0.0825 -0.7889 -0.2268 0.0966 fixed
count_birth_order2/5 -0.09482 0.08333 -1.138 -0.2581 0.06852 fixed
count_birth_order3/5 -0.1679 0.0795 -2.112 -0.3238 -0.01212 fixed
count_birth_order4/5 -0.2076 0.07996 -2.596 -0.3643 -0.05089 fixed
count_birth_order5/5 -0.04856 0.07683 -0.6321 -0.1992 0.102 fixed
count_birth_order1/5+ -0.175 0.08175 -2.141 -0.3352 -0.01479 fixed
count_birth_order2/5+ -0.1714 0.09119 -1.88 -0.3502 0.007309 fixed
count_birth_order3/5+ -0.1815 0.07965 -2.278 -0.3376 -0.02535 fixed
count_birth_order4/5+ -0.2455 0.0742 -3.309 -0.3909 -0.1001 fixed
count_birth_order5/5+ -0.2204 0.0712 -3.095 -0.3599 -0.08083 fixed
count_birth_order5+/5+ -0.181 0.0521 -3.475 -0.2831 -0.07891 fixed
sd_(Intercept).mother_pidlink 0.5413 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7092 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 13795 13868 -6887 13773 NA NA NA
12 13796 13875 -6886 13772 1.298 1 0.2546
16 13800 13905 -6884 13768 4.17 4 0.3834
26 13809 13980 -6879 13757 10.89 10 0.3661

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -8.141 0.9673 -8.417 -10.04 -6.246 fixed
poly(age, 3, raw = TRUE)1 0.9223 0.1445 6.383 0.6391 1.206 fixed
poly(age, 3, raw = TRUE)2 -0.03139 0.007027 -4.468 -0.04517 -0.01762 fixed
poly(age, 3, raw = TRUE)3 0.0003331 0.0001114 2.989 0.0001147 0.0005515 fixed
male -0.08552 0.02357 -3.628 -0.1317 -0.03932 fixed
sibling_count3 0.03226 0.03609 0.8939 -0.03847 0.103 fixed
sibling_count4 -0.07524 0.04109 -1.831 -0.1558 0.005297 fixed
sibling_count5 -0.184 0.05138 -3.58 -0.2847 -0.08326 fixed
sibling_count5+ -0.2158 0.04538 -4.756 -0.3048 -0.1269 fixed
sd_(Intercept).mother_pidlink 0.5446 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7117 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -8.252 0.9682 -8.523 -10.15 -6.355 fixed
birth_order 0.01932 0.009236 2.092 0.001217 0.03742 fixed
poly(age, 3, raw = TRUE)1 0.9344 0.1445 6.465 0.6511 1.218 fixed
poly(age, 3, raw = TRUE)2 -0.03198 0.007028 -4.551 -0.04576 -0.01821 fixed
poly(age, 3, raw = TRUE)3 0.0003432 0.0001115 3.078 0.0001247 0.0005617 fixed
male -0.08584 0.02356 -3.643 -0.132 -0.03966 fixed
sibling_count3 0.02289 0.03636 0.6295 -0.04838 0.09417 fixed
sibling_count4 -0.09874 0.0426 -2.318 -0.1822 -0.01526 fixed
sibling_count5 -0.2229 0.05465 -4.079 -0.33 -0.1158 fixed
sibling_count5+ -0.3012 0.06103 -4.935 -0.4208 -0.1816 fixed
sd_(Intercept).mother_pidlink 0.5451 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.711 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -8.254 0.9693 -8.516 -10.15 -6.355 fixed
poly(age, 3, raw = TRUE)1 0.9354 0.1447 6.465 0.6519 1.219 fixed
poly(age, 3, raw = TRUE)2 -0.032 0.007035 -4.548 -0.04579 -0.01821 fixed
poly(age, 3, raw = TRUE)3 0.0003429 0.0001116 3.073 0.0001242 0.0005616 fixed
male -0.08579 0.02358 -3.639 -0.132 -0.03958 fixed
sibling_count3 0.01426 0.03717 0.3837 -0.0586 0.08712 fixed
sibling_count4 -0.1023 0.04469 -2.29 -0.1899 -0.01475 fixed
sibling_count5 -0.2147 0.05797 -3.704 -0.3284 -0.1011 fixed
sibling_count5+ -0.299 0.06396 -4.675 -0.4244 -0.1737 fixed
birth_order_nonlinear2 0.04838 0.02842 1.703 -0.007317 0.1041 fixed
birth_order_nonlinear3 0.07743 0.0376 2.059 0.003735 0.1511 fixed
birth_order_nonlinear4 0.04621 0.04966 0.9307 -0.05111 0.1435 fixed
birth_order_nonlinear5 0.04818 0.06612 0.7287 -0.08142 0.1778 fixed
birth_order_nonlinear5+ 0.1435 0.07113 2.018 0.004106 0.2829 fixed
sd_(Intercept).mother_pidlink 0.5447 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7113 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -8.273 0.9709 -8.521 -10.18 -6.37 fixed
poly(age, 3, raw = TRUE)1 0.938 0.145 6.471 0.6539 1.222 fixed
poly(age, 3, raw = TRUE)2 -0.03211 0.007048 -4.555 -0.04592 -0.01829 fixed
poly(age, 3, raw = TRUE)3 0.0003444 0.0001118 3.081 0.0001253 0.0005636 fixed
male -0.08458 0.02362 -3.581 -0.1309 -0.03828 fixed
count_birth_order2/2 0.04107 0.04564 0.8999 -0.04838 0.1305 fixed
count_birth_order1/3 0.002483 0.04517 0.05496 -0.08605 0.09102 fixed
count_birth_order2/3 0.08983 0.04961 1.811 -0.0074 0.1871 fixed
count_birth_order3/3 0.07019 0.05485 1.28 -0.03731 0.1777 fixed
count_birth_order1/4 -0.1101 0.06158 -1.789 -0.2308 0.01056 fixed
count_birth_order2/4 -0.1067 0.06099 -1.75 -0.2263 0.01283 fixed
count_birth_order3/4 -0.04312 0.06324 -0.6819 -0.1671 0.08082 fixed
count_birth_order4/4 0.008634 0.06295 0.1372 -0.1147 0.132 fixed
count_birth_order1/5 -0.1416 0.09296 -1.523 -0.3238 0.04063 fixed
count_birth_order2/5 -0.2404 0.1024 -2.348 -0.4412 -0.0397 fixed
count_birth_order3/5 -0.1764 0.08923 -1.977 -0.3513 -0.001481 fixed
count_birth_order4/5 -0.229 0.08433 -2.716 -0.3943 -0.06374 fixed
count_birth_order5/5 -0.1126 0.0824 -1.367 -0.2741 0.04886 fixed
count_birth_order1/5+ -0.3354 0.1144 -2.933 -0.5596 -0.1113 fixed
count_birth_order2/5+ -0.159 0.1131 -1.406 -0.3807 0.0627 fixed
count_birth_order3/5+ -0.06401 0.09685 -0.6609 -0.2538 0.1258 fixed
count_birth_order4/5+ -0.339 0.08979 -3.776 -0.515 -0.163 fixed
count_birth_order5/5+ -0.309 0.08007 -3.859 -0.4659 -0.1521 fixed
count_birth_order5+/5+ -0.1624 0.05637 -2.881 -0.2729 -0.0519 fixed
sd_(Intercept).mother_pidlink 0.5433 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7118 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 13372 13444 -6675 13350 NA NA NA
12 13369 13448 -6673 13345 4.379 1 0.03639
16 13374 13479 -6671 13342 3.268 4 0.514
26 13381 13552 -6665 13329 12.8 10 0.2351

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Raven 2015 old

birthorder <- birthorder %>% mutate(outcome = raven_2015_old)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.2589 0.1483 -1.745 -0.5497 0.03184 fixed
poly(age, 3, raw = TRUE)1 0.05817 0.01424 4.085 0.03026 0.08608 fixed
poly(age, 3, raw = TRUE)2 -0.001978 0.0004199 -4.709 -0.002801 -0.001155 fixed
poly(age, 3, raw = TRUE)3 0.00001287 0.000003875 3.32 0.00000527 0.00002046 fixed
male 0.1522 0.01535 9.916 0.1221 0.1823 fixed
sibling_count3 0.03526 0.0328 1.075 -0.02903 0.09955 fixed
sibling_count4 -0.001169 0.03399 -0.03439 -0.06778 0.06545 fixed
sibling_count5 0.03681 0.03559 1.034 -0.03294 0.1066 fixed
sibling_count5+ -0.1097 0.02774 -3.955 -0.1641 -0.05535 fixed
sd_(Intercept).mother_pidlink 0.4621 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8175 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.2593 0.1483 -1.748 -0.5501 0.0314 fixed
birth_order -0.006107 0.003326 -1.836 -0.01263 0.0004111 fixed
poly(age, 3, raw = TRUE)1 0.06006 0.01428 4.207 0.03208 0.08804 fixed
poly(age, 3, raw = TRUE)2 -0.002051 0.0004218 -4.862 -0.002877 -0.001224 fixed
poly(age, 3, raw = TRUE)3 0.00001355 0.000003893 3.482 0.000005924 0.00002118 fixed
male 0.1523 0.01535 9.924 0.1222 0.1824 fixed
sibling_count3 0.03641 0.0328 1.11 -0.02788 0.1007 fixed
sibling_count4 0.00274 0.03404 0.0805 -0.06398 0.06946 fixed
sibling_count5 0.04392 0.03578 1.227 -0.02622 0.114 fixed
sibling_count5+ -0.08692 0.03038 -2.861 -0.1465 -0.02738 fixed
sd_(Intercept).mother_pidlink 0.4614 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8177 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.2533 0.1487 -1.703 -0.5447 0.03815 fixed
poly(age, 3, raw = TRUE)1 0.05911 0.01428 4.14 0.03113 0.08709 fixed
poly(age, 3, raw = TRUE)2 -0.002023 0.0004217 -4.798 -0.00285 -0.001197 fixed
poly(age, 3, raw = TRUE)3 0.00001335 0.000003894 3.429 0.000005722 0.00002099 fixed
male 0.1521 0.01535 9.913 0.122 0.1822 fixed
sibling_count3 0.02663 0.03319 0.8024 -0.03841 0.09167 fixed
sibling_count4 -0.00795 0.03484 -0.2282 -0.07624 0.06034 fixed
sibling_count5 0.02496 0.03693 0.6759 -0.04742 0.09734 fixed
sibling_count5+ -0.1004 0.0317 -3.166 -0.1625 -0.03823 fixed
birth_order_nonlinear2 -0.01635 0.02205 -0.7413 -0.05957 0.02687 fixed
birth_order_nonlinear3 0.03193 0.02594 1.231 -0.01891 0.08278 fixed
birth_order_nonlinear4 -0.01247 0.02955 -0.4222 -0.07039 0.04544 fixed
birth_order_nonlinear5 0.02924 0.03368 0.8682 -0.03677 0.09525 fixed
birth_order_nonlinear5+ -0.04653 0.02843 -1.636 -0.1023 0.009197 fixed
sd_(Intercept).mother_pidlink 0.4618 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8174 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.23 0.1493 -1.54 -0.5227 0.06267 fixed
poly(age, 3, raw = TRUE)1 0.05915 0.01428 4.142 0.03116 0.08713 fixed
poly(age, 3, raw = TRUE)2 -0.002015 0.0004219 -4.777 -0.002842 -0.001188 fixed
poly(age, 3, raw = TRUE)3 0.0000132 0.000003896 3.389 0.000005569 0.00002084 fixed
male 0.1523 0.01535 9.921 0.1222 0.1824 fixed
count_birth_order2/2 -0.08968 0.04298 -2.087 -0.1739 -0.005445 fixed
count_birth_order1/3 0.001331 0.04216 0.03157 -0.0813 0.08396 fixed
count_birth_order2/3 -0.01655 0.04682 -0.3534 -0.1083 0.07523 fixed
count_birth_order3/3 0.02862 0.05232 0.547 -0.07393 0.1312 fixed
count_birth_order1/4 -0.05712 0.04776 -1.196 -0.1507 0.03649 fixed
count_birth_order2/4 -0.00643 0.0504 -0.1276 -0.1052 0.09234 fixed
count_birth_order3/4 -0.003269 0.0545 -0.05998 -0.1101 0.1035 fixed
count_birth_order4/4 -0.07443 0.05735 -1.298 -0.1868 0.03796 fixed
count_birth_order1/5 -0.06156 0.05424 -1.135 -0.1679 0.04474 fixed
count_birth_order2/5 0.001736 0.05671 0.03062 -0.1094 0.1129 fixed
count_birth_order3/5 0.04451 0.05839 0.7622 -0.06993 0.1589 fixed
count_birth_order4/5 -0.03346 0.06165 -0.5427 -0.1543 0.08738 fixed
count_birth_order5/5 0.09651 0.06314 1.529 -0.02724 0.2203 fixed
count_birth_order1/5+ -0.1224 0.04362 -2.805 -0.2078 -0.03686 fixed
count_birth_order2/5+ -0.1417 0.04493 -3.153 -0.2297 -0.05362 fixed
count_birth_order3/5+ -0.1007 0.04394 -2.291 -0.1868 -0.01454 fixed
count_birth_order4/5+ -0.1217 0.04315 -2.82 -0.2063 -0.03712 fixed
count_birth_order5/5+ -0.1216 0.04345 -2.798 -0.2067 -0.03643 fixed
count_birth_order5+/5+ -0.174 0.0348 -5 -0.2422 -0.1058 fixed
sd_(Intercept).mother_pidlink 0.4612 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8177 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 37283 37366 -18631 37261 NA NA NA
12 37282 37372 -18629 37258 3.375 1 0.06621
16 37283 37404 -18625 37251 6.817 4 0.1459
26 37293 37489 -18621 37241 9.765 10 0.4614

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5241 0.3667 -1.429 -1.243 0.1946 fixed
poly(age, 3, raw = TRUE)1 0.09035 0.04162 2.171 0.008782 0.1719 fixed
poly(age, 3, raw = TRUE)2 -0.002746 0.001486 -1.849 -0.005658 0.0001652 fixed
poly(age, 3, raw = TRUE)3 0.00002114 0.00001683 1.256 -0.00001185 0.00005412 fixed
male 0.09392 0.02139 4.392 0.052 0.1358 fixed
sibling_count3 0.0296 0.03472 0.8526 -0.03845 0.09766 fixed
sibling_count4 -0.0397 0.03752 -1.058 -0.1132 0.03383 fixed
sibling_count5 -0.08843 0.04313 -2.05 -0.173 -0.003891 fixed
sibling_count5+ -0.1438 0.03792 -3.792 -0.2181 -0.06949 fixed
sd_(Intercept).mother_pidlink 0.3938 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7424 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5258 0.3668 -1.434 -1.245 0.193 fixed
birth_order 0.002299 0.007146 0.3217 -0.01171 0.0163 fixed
poly(age, 3, raw = TRUE)1 0.09021 0.04162 2.167 0.008632 0.1718 fixed
poly(age, 3, raw = TRUE)2 -0.002745 0.001486 -1.848 -0.005657 0.000167 fixed
poly(age, 3, raw = TRUE)3 0.00002121 0.00001683 1.26 -0.00001178 0.0000542 fixed
male 0.09381 0.02139 4.386 0.05188 0.1357 fixed
sibling_count3 0.02845 0.03491 0.815 -0.03997 0.09687 fixed
sibling_count4 -0.0424 0.03845 -1.103 -0.1178 0.03297 fixed
sibling_count5 -0.09284 0.04527 -2.051 -0.1816 -0.004117 fixed
sibling_count5+ -0.1526 0.04681 -3.261 -0.2444 -0.06089 fixed
sd_(Intercept).mother_pidlink 0.394 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7423 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5665 0.3675 -1.541 -1.287 0.1539 fixed
poly(age, 3, raw = TRUE)1 0.09346 0.04166 2.243 0.01181 0.1751 fixed
poly(age, 3, raw = TRUE)2 -0.00286 0.001487 -1.923 -0.005774 0.00005512 fixed
poly(age, 3, raw = TRUE)3 0.00002249 0.00001685 1.335 -0.00001053 0.00005552 fixed
male 0.09389 0.02139 4.389 0.05196 0.1358 fixed
sibling_count3 0.01664 0.03555 0.468 -0.05304 0.08633 fixed
sibling_count4 -0.05847 0.03981 -1.469 -0.1365 0.01957 fixed
sibling_count5 -0.1133 0.04732 -2.394 -0.2061 -0.02056 fixed
sibling_count5+ -0.1617 0.04806 -3.365 -0.2559 -0.06754 fixed
birth_order_nonlinear2 0.04528 0.02729 1.659 -0.008212 0.09877 fixed
birth_order_nonlinear3 0.0584 0.03378 1.729 -0.007813 0.1246 fixed
birth_order_nonlinear4 0.03984 0.04188 0.9514 -0.04224 0.1219 fixed
birth_order_nonlinear5 0.05613 0.05225 1.074 -0.04627 0.1585 fixed
birth_order_nonlinear5+ 0.01605 0.05306 0.3025 -0.08795 0.12 fixed
sd_(Intercept).mother_pidlink 0.3928 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7428 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5931 0.3685 -1.61 -1.315 0.1291 fixed
poly(age, 3, raw = TRUE)1 0.09665 0.04176 2.314 0.0148 0.1785 fixed
poly(age, 3, raw = TRUE)2 -0.002961 0.001491 -1.986 -0.005883 -0.00003867 fixed
poly(age, 3, raw = TRUE)3 0.00002349 0.0000169 1.39 -0.000009634 0.00005661 fixed
male 0.09428 0.02141 4.403 0.05231 0.1363 fixed
count_birth_order2/2 0.03312 0.04956 0.6683 -0.06401 0.1302 fixed
count_birth_order1/3 -0.01387 0.04484 -0.3094 -0.1018 0.07402 fixed
count_birth_order2/3 0.08776 0.04873 1.801 -0.007741 0.1833 fixed
count_birth_order3/3 0.08144 0.0544 1.497 -0.02519 0.1881 fixed
count_birth_order1/4 -0.06891 0.05445 -1.265 -0.1756 0.03781 fixed
count_birth_order2/4 -0.02435 0.05634 -0.4322 -0.1348 0.08608 fixed
count_birth_order3/4 0.006464 0.05937 0.1089 -0.1099 0.1228 fixed
count_birth_order4/4 -0.0157 0.06178 -0.2541 -0.1368 0.1054 fixed
count_birth_order1/5 -0.05962 0.07384 -0.8075 -0.2043 0.0851 fixed
count_birth_order2/5 -0.04247 0.07901 -0.5376 -0.1973 0.1124 fixed
count_birth_order3/5 -0.05223 0.07429 -0.703 -0.1978 0.09337 fixed
count_birth_order4/5 -0.1066 0.07177 -1.486 -0.2473 0.03402 fixed
count_birth_order5/5 -0.1237 0.07425 -1.666 -0.2692 0.02181 fixed
count_birth_order1/5+ -0.1206 0.07347 -1.642 -0.2646 0.02337 fixed
count_birth_order2/5+ -0.1956 0.07265 -2.693 -0.338 -0.05325 fixed
count_birth_order3/5+ -0.1502 0.07269 -2.066 -0.2926 -0.007687 fixed
count_birth_order4/5+ -0.1115 0.06833 -1.631 -0.2454 0.02246 fixed
count_birth_order5/5+ -0.06799 0.06468 -1.051 -0.1948 0.05878 fixed
count_birth_order5+/5+ -0.1493 0.0495 -3.016 -0.2463 -0.05226 fixed
sd_(Intercept).mother_pidlink 0.3924 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7433 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 14652 14726 -7315 14630 NA NA NA
12 14654 14734 -7315 14630 0.1023 1 0.7491
16 14657 14764 -7313 14625 4.598 4 0.3311
26 14671 14844 -7309 14619 6.858 10 0.7388

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.4997 0.3654 -1.367 -1.216 0.2166 fixed
poly(age, 3, raw = TRUE)1 0.08706 0.04149 2.098 0.005739 0.1684 fixed
poly(age, 3, raw = TRUE)2 -0.002669 0.001481 -1.801 -0.005572 0.000235 fixed
poly(age, 3, raw = TRUE)3 0.0000203 0.00001679 1.209 -0.0000126 0.0000532 fixed
male 0.09351 0.02129 4.392 0.05178 0.1352 fixed
sibling_count3 0.03147 0.0375 0.8392 -0.04204 0.105 fixed
sibling_count4 -0.009716 0.03964 -0.2451 -0.08741 0.06797 fixed
sibling_count5 -0.01839 0.04258 -0.4318 -0.1019 0.06508 fixed
sibling_count5+ -0.0805 0.03731 -2.157 -0.1536 -0.007369 fixed
sd_(Intercept).mother_pidlink 0.3954 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7416 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.4938 0.3655 -1.351 -1.21 0.2226 fixed
birth_order -0.005347 0.006257 -0.8546 -0.01761 0.006915 fixed
poly(age, 3, raw = TRUE)1 0.08718 0.0415 2.101 0.005853 0.1685 fixed
poly(age, 3, raw = TRUE)2 -0.002665 0.001482 -1.799 -0.005569 0.000239 fixed
poly(age, 3, raw = TRUE)3 0.00002004 0.00001679 1.193 -0.00001288 0.00005295 fixed
male 0.09375 0.0213 4.402 0.05201 0.1355 fixed
sibling_count3 0.03414 0.03762 0.9075 -0.0396 0.1079 fixed
sibling_count4 -0.003681 0.04025 -0.09145 -0.08257 0.07521 fixed
sibling_count5 -0.008885 0.044 -0.2019 -0.09513 0.07736 fixed
sibling_count5+ -0.06066 0.04394 -1.381 -0.1468 0.02546 fixed
sd_(Intercept).mother_pidlink 0.3946 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7419 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5303 0.3662 -1.448 -1.248 0.1875 fixed
poly(age, 3, raw = TRUE)1 0.08914 0.04153 2.146 0.00774 0.1705 fixed
poly(age, 3, raw = TRUE)2 -0.002734 0.001483 -1.844 -0.005641 0.0001724 fixed
poly(age, 3, raw = TRUE)3 0.00002084 0.00001681 1.24 -0.00001211 0.00005378 fixed
male 0.09355 0.0213 4.391 0.05179 0.1353 fixed
sibling_count3 0.02214 0.03824 0.579 -0.05281 0.09708 fixed
sibling_count4 -0.02317 0.04153 -0.558 -0.1046 0.05823 fixed
sibling_count5 -0.03049 0.04591 -0.6641 -0.1205 0.05949 fixed
sibling_count5+ -0.07221 0.04516 -1.599 -0.1607 0.0163 fixed
birth_order_nonlinear2 0.036 0.02783 1.294 -0.01855 0.09056 fixed
birth_order_nonlinear3 0.04407 0.03368 1.309 -0.02194 0.1101 fixed
birth_order_nonlinear4 0.03472 0.04065 0.8542 -0.04494 0.1144 fixed
birth_order_nonlinear5 0.01614 0.04974 0.3245 -0.08135 0.1136 fixed
birth_order_nonlinear5+ -0.03226 0.04754 -0.6786 -0.1254 0.06092 fixed
sd_(Intercept).mother_pidlink 0.3933 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7425 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.527 0.3668 -1.437 -1.246 0.192 fixed
poly(age, 3, raw = TRUE)1 0.08859 0.04158 2.131 0.007097 0.1701 fixed
poly(age, 3, raw = TRUE)2 -0.002699 0.001485 -1.817 -0.005609 0.0002117 fixed
poly(age, 3, raw = TRUE)3 0.00002026 0.00001683 1.203 -0.00001273 0.00005325 fixed
male 0.09351 0.0213 4.389 0.05176 0.1353 fixed
count_birth_order2/2 0.03047 0.05428 0.5614 -0.07591 0.1369 fixed
count_birth_order1/3 -0.005049 0.0485 -0.1041 -0.1001 0.09001 fixed
count_birth_order2/3 0.07102 0.05237 1.356 -0.03162 0.1737 fixed
count_birth_order3/3 0.09355 0.05888 1.589 -0.02185 0.209 fixed
count_birth_order1/4 -0.07743 0.05693 -1.36 -0.189 0.03414 fixed
count_birth_order2/4 0.04346 0.05822 0.7466 -0.07064 0.1576 fixed
count_birth_order3/4 0.04852 0.0635 0.7642 -0.07593 0.173 fixed
count_birth_order4/4 0.01437 0.06557 0.2191 -0.1142 0.1429 fixed
count_birth_order1/5 0.02045 0.06753 0.3029 -0.1119 0.1528 fixed
count_birth_order2/5 0.05005 0.07271 0.6884 -0.09246 0.1926 fixed
count_birth_order3/5 -0.02403 0.07055 -0.3406 -0.1623 0.1143 fixed
count_birth_order4/5 -0.03049 0.07305 -0.4174 -0.1737 0.1127 fixed
count_birth_order5/5 -0.06213 0.07291 -0.8522 -0.205 0.08076 fixed
count_birth_order1/5+ 0.008191 0.06447 0.127 -0.1182 0.1345 fixed
count_birth_order2/5+ -0.146 0.06705 -2.177 -0.2774 -0.01458 fixed
count_birth_order3/5+ -0.06925 0.06562 -1.055 -0.1979 0.05935 fixed
count_birth_order4/5+ -0.02285 0.06336 -0.3607 -0.147 0.1013 fixed
count_birth_order5/5+ -0.02685 0.06489 -0.4137 -0.154 0.1003 fixed
count_birth_order5+/5+ -0.1068 0.04849 -2.204 -0.2019 -0.01182 fixed
sd_(Intercept).mother_pidlink 0.3949 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7416 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 14783 14856 -7380 14761 NA NA NA
12 14784 14864 -7380 14760 0.7334 1 0.3918
16 14788 14895 -7378 14756 3.844 4 0.4275
26 14795 14969 -7371 14743 13.4 10 0.202

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5761 0.3711 -1.552 -1.303 0.1513 fixed
poly(age, 3, raw = TRUE)1 0.09716 0.04213 2.306 0.01458 0.1797 fixed
poly(age, 3, raw = TRUE)2 -0.003018 0.001504 -2.006 -0.005966 -0.00006911 fixed
poly(age, 3, raw = TRUE)3 0.00002451 0.00001705 1.437 -0.000008914 0.00005793 fixed
male 0.09125 0.02161 4.222 0.04889 0.1336 fixed
sibling_count3 0.01999 0.03422 0.5842 -0.04708 0.08707 fixed
sibling_count4 -0.03413 0.03724 -0.9165 -0.1071 0.03886 fixed
sibling_count5 -0.08933 0.0442 -2.021 -0.176 -0.002698 fixed
sibling_count5+ -0.1448 0.03831 -3.779 -0.2199 -0.06971 fixed
sd_(Intercept).mother_pidlink 0.3902 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7437 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5775 0.3712 -1.556 -1.305 0.1499 fixed
birth_order 0.002238 0.007356 0.3042 -0.01218 0.01666 fixed
poly(age, 3, raw = TRUE)1 0.097 0.04214 2.302 0.01441 0.1796 fixed
poly(age, 3, raw = TRUE)2 -0.003016 0.001505 -2.004 -0.005964 -0.00006674 fixed
poly(age, 3, raw = TRUE)3 0.00002458 0.00001706 1.441 -0.00000885 0.00005801 fixed
male 0.09118 0.02162 4.218 0.04882 0.1336 fixed
sibling_count3 0.01887 0.03442 0.5481 -0.0486 0.08634 fixed
sibling_count4 -0.03673 0.03821 -0.9612 -0.1116 0.03816 fixed
sibling_count5 -0.09345 0.04623 -2.022 -0.184 -0.002845 fixed
sibling_count5+ -0.1533 0.04742 -3.233 -0.2462 -0.06036 fixed
sd_(Intercept).mother_pidlink 0.3905 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7437 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.6256 0.3718 -1.683 -1.354 0.1032 fixed
poly(age, 3, raw = TRUE)1 0.1004 0.04217 2.382 0.0178 0.1831 fixed
poly(age, 3, raw = TRUE)2 -0.003136 0.001506 -2.083 -0.006087 -0.000185 fixed
poly(age, 3, raw = TRUE)3 0.00002589 0.00001707 1.517 -0.000007558 0.00005935 fixed
male 0.0914 0.02162 4.228 0.04903 0.1338 fixed
sibling_count3 0.008164 0.03508 0.2327 -0.06059 0.07692 fixed
sibling_count4 -0.05403 0.03961 -1.364 -0.1317 0.02361 fixed
sibling_count5 -0.1078 0.04814 -2.239 -0.2021 -0.01343 fixed
sibling_count5+ -0.1591 0.04873 -3.265 -0.2546 -0.0636 fixed
birth_order_nonlinear2 0.06079 0.02726 2.23 0.007365 0.1142 fixed
birth_order_nonlinear3 0.05707 0.03379 1.689 -0.00916 0.1233 fixed
birth_order_nonlinear4 0.05315 0.04308 1.234 -0.03129 0.1376 fixed
birth_order_nonlinear5 0.02966 0.05454 0.5438 -0.07724 0.1366 fixed
birth_order_nonlinear5+ 0.02338 0.05469 0.4275 -0.08381 0.1306 fixed
sd_(Intercept).mother_pidlink 0.3892 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7441 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.6437 0.3729 -1.726 -1.375 0.08717 fixed
poly(age, 3, raw = TRUE)1 0.1026 0.04228 2.427 0.01973 0.1855 fixed
poly(age, 3, raw = TRUE)2 -0.003204 0.00151 -2.122 -0.006164 -0.0002446 fixed
poly(age, 3, raw = TRUE)3 0.00002657 0.00001712 1.552 -0.000006992 0.00006013 fixed
male 0.09147 0.02164 4.227 0.04905 0.1339 fixed
count_birth_order2/2 0.05327 0.04822 1.105 -0.04123 0.1478 fixed
count_birth_order1/3 -0.01454 0.04425 -0.3286 -0.1013 0.07219 fixed
count_birth_order2/3 0.09972 0.04866 2.049 0.004357 0.1951 fixed
count_birth_order3/3 0.05742 0.05341 1.075 -0.04726 0.1621 fixed
count_birth_order1/4 -0.0668 0.05467 -1.222 -0.174 0.04036 fixed
count_birth_order2/4 -0.003471 0.05641 -0.06154 -0.114 0.1071 fixed
count_birth_order3/4 0.0155 0.05882 0.2634 -0.0998 0.1308 fixed
count_birth_order4/4 0.004197 0.06202 0.06768 -0.1174 0.1257 fixed
count_birth_order1/5 -0.05474 0.07389 -0.7408 -0.1996 0.09008 fixed
count_birth_order2/5 -0.04276 0.08163 -0.5239 -0.2028 0.1172 fixed
count_birth_order3/5 -0.04845 0.07764 -0.624 -0.2006 0.1037 fixed
count_birth_order4/5 -0.07973 0.07529 -1.059 -0.2273 0.06784 fixed
count_birth_order5/5 -0.1367 0.07946 -1.72 -0.2924 0.01905 fixed
count_birth_order1/5+ -0.1208 0.0752 -1.607 -0.2682 0.02654 fixed
count_birth_order2/5+ -0.1781 0.07448 -2.391 -0.3241 -0.03211 fixed
count_birth_order3/5+ -0.1208 0.07353 -1.643 -0.2649 0.02331 fixed
count_birth_order4/5+ -0.1009 0.07186 -1.404 -0.2417 0.03994 fixed
count_birth_order5/5+ -0.09842 0.06622 -1.486 -0.2282 0.03136 fixed
count_birth_order5+/5+ -0.1381 0.0504 -2.74 -0.2369 -0.03932 fixed
sd_(Intercept).mother_pidlink 0.3886 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7447 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 14356 14429 -7167 14334 NA NA NA
12 14358 14438 -7167 14334 0.09157 1 0.7622
16 14360 14467 -7164 14328 6.1 4 0.1918
26 14375 14548 -7161 14323 5.282 10 0.8716

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Numeracy 2015 old

birthorder <- birthorder %>% mutate(outcome = math_2015_old)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.3193 0.2143 1.49 -0.1007 0.7392 fixed
poly(age, 3, raw = TRUE)1 0.007294 0.02199 0.3316 -0.03581 0.0504 fixed
poly(age, 3, raw = TRUE)2 -0.0005758 0.0007025 -0.8196 -0.001953 0.0008011 fixed
poly(age, 3, raw = TRUE)3 0.000004245 0.000007055 0.6017 -0.000009582 0.00001807 fixed
male -0.07804 0.01633 -4.778 -0.11 -0.04602 fixed
sibling_count3 0.03562 0.03434 1.037 -0.03169 0.1029 fixed
sibling_count4 -0.04503 0.03547 -1.27 -0.1145 0.02448 fixed
sibling_count5 -0.01284 0.03704 -0.3466 -0.08544 0.05976 fixed
sibling_count5+ -0.1715 0.02895 -5.922 -0.2282 -0.1147 fixed
sd_(Intercept).mother_pidlink 0.4447 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.881 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.3021 0.2145 1.409 -0.1182 0.7225 fixed
birth_order -0.006092 0.003514 -1.734 -0.01298 0.0007958 fixed
poly(age, 3, raw = TRUE)1 0.01092 0.02209 0.4941 -0.03238 0.05422 fixed
poly(age, 3, raw = TRUE)2 -0.0007057 0.0007065 -0.9988 -0.00209 0.0006791 fixed
poly(age, 3, raw = TRUE)3 0.000005524 0.000007093 0.7787 -0.000008379 0.00001943 fixed
male -0.07783 0.01633 -4.765 -0.1098 -0.04582 fixed
sibling_count3 0.03689 0.03434 1.074 -0.03042 0.1042 fixed
sibling_count4 -0.04087 0.03554 -1.15 -0.1105 0.02879 fixed
sibling_count5 -0.005543 0.03727 -0.1487 -0.07859 0.0675 fixed
sibling_count5+ -0.1485 0.03184 -4.664 -0.2109 -0.08609 fixed
sd_(Intercept).mother_pidlink 0.4442 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8811 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.3361 0.2158 1.558 -0.0868 0.7591 fixed
poly(age, 3, raw = TRUE)1 0.006561 0.02214 0.2963 -0.03683 0.04995 fixed
poly(age, 3, raw = TRUE)2 -0.0005546 0.0007086 -0.7827 -0.001943 0.0008343 fixed
poly(age, 3, raw = TRUE)3 0.000004035 0.000007122 0.5665 -0.000009925 0.000018 fixed
male -0.0781 0.01634 -4.781 -0.1101 -0.04609 fixed
sibling_count3 0.03465 0.03477 0.9965 -0.0335 0.1028 fixed
sibling_count4 -0.05003 0.03642 -1.374 -0.1214 0.02136 fixed
sibling_count5 -0.02144 0.03854 -0.5564 -0.09697 0.05409 fixed
sibling_count5+ -0.1731 0.03329 -5.201 -0.2384 -0.1079 fixed
birth_order_nonlinear2 -0.02299 0.02364 -0.9727 -0.06933 0.02334 fixed
birth_order_nonlinear3 -0.002078 0.02775 -0.07488 -0.05647 0.05232 fixed
birth_order_nonlinear4 0.01697 0.03155 0.5378 -0.04487 0.07881 fixed
birth_order_nonlinear5 0.01641 0.03594 0.4564 -0.05404 0.08685 fixed
birth_order_nonlinear5+ -0.01964 0.03019 -0.6505 -0.0788 0.03953 fixed
sd_(Intercept).mother_pidlink 0.4443 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8812 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.3273 0.217 1.509 -0.09794 0.7526 fixed
poly(age, 3, raw = TRUE)1 0.006943 0.02219 0.3129 -0.03654 0.05043 fixed
poly(age, 3, raw = TRUE)2 -0.0005669 0.0007105 -0.7979 -0.00196 0.0008257 fixed
poly(age, 3, raw = TRUE)3 0.000004164 0.000007146 0.5828 -0.000009841 0.00001817 fixed
male -0.07788 0.01634 -4.766 -0.1099 -0.04585 fixed
count_birth_order2/2 -0.00982 0.04618 -0.2127 -0.1003 0.08068 fixed
count_birth_order1/3 0.05031 0.04467 1.126 -0.03724 0.1379 fixed
count_birth_order2/3 0.01198 0.04969 0.2411 -0.08541 0.1094 fixed
count_birth_order3/3 0.0213 0.05561 0.3831 -0.08768 0.1303 fixed
count_birth_order1/4 -0.05016 0.05062 -0.9909 -0.1494 0.04906 fixed
count_birth_order2/4 -0.06037 0.05338 -1.131 -0.165 0.04425 fixed
count_birth_order3/4 -0.03141 0.05772 -0.5442 -0.1445 0.08172 fixed
count_birth_order4/4 -0.04884 0.06074 -0.8041 -0.1679 0.0702 fixed
count_birth_order1/5 -0.05318 0.05742 -0.9263 -0.1657 0.05935 fixed
count_birth_order2/5 -0.08221 0.06012 -1.367 -0.2 0.03562 fixed
count_birth_order3/5 0.01386 0.06183 0.2241 -0.1073 0.1351 fixed
count_birth_order4/5 0.01906 0.06531 0.2918 -0.1089 0.1471 fixed
count_birth_order5/5 0.05706 0.06688 0.8531 -0.07403 0.1881 fixed
count_birth_order1/5+ -0.1509 0.04641 -3.251 -0.2418 -0.05991 fixed
count_birth_order2/5+ -0.1789 0.04772 -3.75 -0.2725 -0.08542 fixed
count_birth_order3/5+ -0.183 0.04667 -3.921 -0.2744 -0.09152 fixed
count_birth_order4/5+ -0.1483 0.04581 -3.238 -0.2381 -0.05853 fixed
count_birth_order5/5+ -0.1708 0.04611 -3.705 -0.2612 -0.08044 fixed
count_birth_order5+/5+ -0.1879 0.03668 -5.123 -0.2598 -0.116 fixed
sd_(Intercept).mother_pidlink 0.444 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8815 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 38695 38778 -19337 38673 NA NA NA
12 38694 38785 -19335 38670 3.008 1 0.08286
16 38702 38823 -19335 38670 0.06628 4 0.9995
26 38718 38914 -19333 38666 4.383 10 0.9284

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.4124 0.4482 -0.92 -1.291 0.4662 fixed
poly(age, 3, raw = TRUE)1 0.09996 0.05087 1.965 0.0002534 0.1997 fixed
poly(age, 3, raw = TRUE)2 -0.003619 0.001816 -1.993 -0.007178 -0.00005981 fixed
poly(age, 3, raw = TRUE)3 0.00003836 0.00002057 1.865 -0.000001958 0.00007868 fixed
male -0.1536 0.02616 -5.869 -0.2048 -0.1023 fixed
sibling_count3 0.01062 0.04193 0.2533 -0.07156 0.0928 fixed
sibling_count4 -0.08796 0.04522 -1.945 -0.1766 0.0006691 fixed
sibling_count5 -0.1347 0.05191 -2.594 -0.2364 -0.03293 fixed
sibling_count5+ -0.2428 0.04561 -5.325 -0.3322 -0.1534 fixed
sd_(Intercept).mother_pidlink 0.4377 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.923 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.4249 0.4482 -0.9481 -1.303 0.4535 fixed
birth_order 0.01582 0.008711 1.817 -0.001249 0.0329 fixed
poly(age, 3, raw = TRUE)1 0.09906 0.05086 1.948 -0.0006253 0.1988 fixed
poly(age, 3, raw = TRUE)2 -0.003611 0.001816 -1.989 -0.007169 -0.00005259 fixed
poly(age, 3, raw = TRUE)3 0.00003891 0.00002057 1.892 -0.000001407 0.00007922 fixed
male -0.1543 0.02616 -5.898 -0.2056 -0.103 fixed
sibling_count3 0.002788 0.04214 0.06616 -0.07981 0.08539 fixed
sibling_count4 -0.1064 0.04634 -2.296 -0.1972 -0.01558 fixed
sibling_count5 -0.1648 0.05449 -3.024 -0.2716 -0.05801 fixed
sibling_count5+ -0.3032 0.05643 -5.374 -0.4138 -0.1926 fixed
sd_(Intercept).mother_pidlink 0.4378 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9227 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.4289 0.4491 -0.955 -1.309 0.4513 fixed
poly(age, 3, raw = TRUE)1 0.1012 0.05091 1.987 0.001384 0.2009 fixed
poly(age, 3, raw = TRUE)2 -0.003688 0.001817 -2.029 -0.00725 -0.0001261 fixed
poly(age, 3, raw = TRUE)3 0.00003982 0.00002059 1.934 -0.0000005415 0.00008018 fixed
male -0.1543 0.02617 -5.895 -0.2055 -0.103 fixed
sibling_count3 -0.006667 0.04296 -0.1552 -0.09088 0.07754 fixed
sibling_count4 -0.1178 0.04805 -2.451 -0.212 -0.02361 fixed
sibling_count5 -0.1912 0.05708 -3.35 -0.3031 -0.07933 fixed
sibling_count5+ -0.317 0.058 -5.465 -0.4306 -0.2033 fixed
birth_order_nonlinear2 0.02081 0.03354 0.6204 -0.04493 0.08655 fixed
birth_order_nonlinear3 0.07339 0.04148 1.769 -0.007901 0.1547 fixed
birth_order_nonlinear4 0.05819 0.05138 1.132 -0.04251 0.1589 fixed
birth_order_nonlinear5 0.1455 0.06415 2.269 0.01979 0.2713 fixed
birth_order_nonlinear5+ 0.09315 0.06484 1.437 -0.03393 0.2202 fixed
sd_(Intercept).mother_pidlink 0.4367 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9232 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.4642 0.4503 -1.031 -1.347 0.4183 fixed
poly(age, 3, raw = TRUE)1 0.1045 0.05104 2.048 0.004513 0.2046 fixed
poly(age, 3, raw = TRUE)2 -0.003813 0.001822 -2.093 -0.007384 -0.0002416 fixed
poly(age, 3, raw = TRUE)3 0.00004129 0.00002065 1.999 0.0000008111 0.00008176 fixed
male -0.1531 0.0262 -5.845 -0.2044 -0.1018 fixed
count_birth_order2/2 0.03911 0.06081 0.6432 -0.08008 0.1583 fixed
count_birth_order1/3 0.004539 0.0546 0.08314 -0.1025 0.1115 fixed
count_birth_order2/3 0.02779 0.05938 0.4679 -0.0886 0.1442 fixed
count_birth_order3/3 0.05161 0.06635 0.7778 -0.07844 0.1816 fixed
count_birth_order1/4 -0.1031 0.06635 -1.553 -0.2331 0.02698 fixed
count_birth_order2/4 -0.131 0.06871 -1.907 -0.2657 0.003622 fixed
count_birth_order3/4 -0.03292 0.07245 -0.4543 -0.1749 0.1091 fixed
count_birth_order4/4 -0.02071 0.0754 -0.2747 -0.1685 0.1271 fixed
count_birth_order1/5 -0.1396 0.09008 -1.55 -0.3162 0.03694 fixed
count_birth_order2/5 -0.1104 0.0965 -1.144 -0.2995 0.07874 fixed
count_birth_order3/5 -0.06116 0.09072 -0.6741 -0.239 0.1167 fixed
count_birth_order4/5 -0.2161 0.08761 -2.466 -0.3878 -0.04436 fixed
count_birth_order5/5 -0.08756 0.09067 -0.9657 -0.2653 0.09014 fixed
count_birth_order1/5+ -0.3657 0.08971 -4.076 -0.5415 -0.1899 fixed
count_birth_order2/5+ -0.3071 0.08873 -3.461 -0.481 -0.1332 fixed
count_birth_order3/5+ -0.2507 0.08883 -2.822 -0.4248 -0.07655 fixed
count_birth_order4/5+ -0.2177 0.0835 -2.607 -0.3813 -0.05401 fixed
count_birth_order5/5+ -0.1333 0.07903 -1.687 -0.2882 0.02158 fixed
count_birth_order5+/5+ -0.2173 0.06012 -3.614 -0.3351 -0.09944 fixed
sd_(Intercept).mother_pidlink 0.4363 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9239 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 17008 17082 -8493 16986 NA NA NA
12 17007 17087 -8492 16983 3.304 1 0.06913
16 17012 17119 -8490 16980 3.482 4 0.4806
26 17027 17200 -8487 16975 4.996 10 0.8914

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.3743 0.4477 -0.8361 -1.252 0.5032 fixed
poly(age, 3, raw = TRUE)1 0.09573 0.05083 1.883 -0.003898 0.1954 fixed
poly(age, 3, raw = TRUE)2 -0.003491 0.001815 -1.924 -0.007049 0.00006595 fixed
poly(age, 3, raw = TRUE)3 0.00003683 0.00002057 1.791 -0.000003482 0.00007714 fixed
male -0.1544 0.02611 -5.913 -0.2056 -0.1032 fixed
sibling_count3 0.02706 0.04542 0.5958 -0.06196 0.1161 fixed
sibling_count4 -0.07306 0.04794 -1.524 -0.167 0.0209 fixed
sibling_count5 -0.08228 0.05143 -1.6 -0.1831 0.01851 fixed
sibling_count5+ -0.1821 0.04509 -4.039 -0.2705 -0.09372 fixed
sd_(Intercept).mother_pidlink 0.4413 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.924 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.3791 0.4478 -0.8465 -1.257 0.4986 fixed
birth_order 0.004275 0.007636 0.5599 -0.01069 0.01924 fixed
poly(age, 3, raw = TRUE)1 0.09565 0.05084 1.882 -0.003984 0.1953 fixed
poly(age, 3, raw = TRUE)2 -0.003495 0.001815 -1.926 -0.007052 0.00006254 fixed
poly(age, 3, raw = TRUE)3 0.00003704 0.00002057 1.801 -0.000003279 0.00007736 fixed
male -0.1546 0.02611 -5.919 -0.2057 -0.1034 fixed
sibling_count3 0.02495 0.04558 0.5475 -0.06438 0.1143 fixed
sibling_count4 -0.07783 0.04869 -1.598 -0.1733 0.0176 fixed
sibling_count5 -0.08982 0.05317 -1.689 -0.194 0.01438 fixed
sibling_count5+ -0.1979 0.05319 -3.721 -0.3021 -0.09364 fixed
sd_(Intercept).mother_pidlink 0.4415 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.924 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.3883 0.4486 -0.8654 -1.268 0.491 fixed
poly(age, 3, raw = TRUE)1 0.09661 0.05088 1.899 -0.003112 0.1963 fixed
poly(age, 3, raw = TRUE)2 -0.003535 0.001817 -1.946 -0.007096 0.00002593 fixed
poly(age, 3, raw = TRUE)3 0.00003765 0.00002059 1.828 -0.000002709 0.00007801 fixed
male -0.1548 0.02612 -5.928 -0.206 -0.1036 fixed
sibling_count3 0.01383 0.04637 0.2983 -0.07706 0.1047 fixed
sibling_count4 -0.09552 0.05032 -1.898 -0.1942 0.003102 fixed
sibling_count5 -0.1103 0.05558 -1.984 -0.2192 -0.001355 fixed
sibling_count5+ -0.2216 0.05474 -4.048 -0.3289 -0.1143 fixed
birth_order_nonlinear2 0.02009 0.03426 0.5864 -0.04706 0.08725 fixed
birth_order_nonlinear3 0.05782 0.04142 1.396 -0.02337 0.139 fixed
birth_order_nonlinear4 0.05083 0.04996 1.018 -0.04708 0.1488 fixed
birth_order_nonlinear5 0.0438 0.06117 0.716 -0.07609 0.1637 fixed
birth_order_nonlinear5+ 0.05985 0.05819 1.029 -0.05419 0.1739 fixed
sd_(Intercept).mother_pidlink 0.4413 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9242 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.4364 0.4495 -0.9708 -1.317 0.4447 fixed
poly(age, 3, raw = TRUE)1 0.1003 0.05096 1.969 0.0004537 0.2002 fixed
poly(age, 3, raw = TRUE)2 -0.003676 0.00182 -2.02 -0.007243 -0.0001092 fixed
poly(age, 3, raw = TRUE)3 0.00003931 0.00002063 1.906 -0.000001122 0.00007975 fixed
male -0.1557 0.02614 -5.956 -0.2069 -0.1044 fixed
count_birth_order2/2 0.07481 0.06677 1.12 -0.05605 0.2057 fixed
count_birth_order1/3 0.06574 0.05921 1.11 -0.05032 0.1818 fixed
count_birth_order2/3 0.007638 0.06399 0.1194 -0.1178 0.1331 fixed
count_birth_order3/3 0.08768 0.07201 1.218 -0.05346 0.2288 fixed
count_birth_order1/4 -0.1036 0.06955 -1.49 -0.24 0.03269 fixed
count_birth_order2/4 -0.03628 0.07118 -0.5097 -0.1758 0.1032 fixed
count_birth_order3/4 -0.02561 0.07771 -0.3295 -0.1779 0.1267 fixed
count_birth_order4/4 -0.006286 0.08024 -0.07833 -0.1636 0.151 fixed
count_birth_order1/5 -0.007333 0.08258 -0.0888 -0.1692 0.1545 fixed
count_birth_order2/5 -0.07043 0.08902 -0.7912 -0.2449 0.104 fixed
count_birth_order3/5 0.01048 0.08637 0.1213 -0.1588 0.1798 fixed
count_birth_order4/5 -0.1567 0.08947 -1.752 -0.3321 0.01864 fixed
count_birth_order5/5 -0.09262 0.08928 -1.037 -0.2676 0.08236 fixed
count_birth_order1/5+ -0.2681 0.07886 -3.399 -0.4226 -0.1135 fixed
count_birth_order2/5+ -0.1808 0.0821 -2.202 -0.3417 -0.01989 fixed
count_birth_order3/5+ -0.1727 0.08036 -2.149 -0.3302 -0.0152 fixed
count_birth_order4/5+ -0.09252 0.0776 -1.192 -0.2446 0.05957 fixed
count_birth_order5/5+ -0.128 0.07952 -1.61 -0.2839 0.02784 fixed
count_birth_order5+/5+ -0.1425 0.05907 -2.413 -0.2583 -0.02678 fixed
sd_(Intercept).mother_pidlink 0.4412 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9243 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 17187 17261 -8583 17165 NA NA NA
12 17189 17269 -8582 17165 0.3133 1 0.5757
16 17195 17302 -8581 17163 2.203 4 0.6986
26 17205 17379 -8577 17153 9.439 10 0.491

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5356 0.4544 -1.179 -1.426 0.355 fixed
poly(age, 3, raw = TRUE)1 0.1138 0.05159 2.206 0.01271 0.215 fixed
poly(age, 3, raw = TRUE)2 -0.004171 0.001842 -2.264 -0.007782 -0.0005606 fixed
poly(age, 3, raw = TRUE)3 0.00004509 0.00002088 2.16 0.000004169 0.00008602 fixed
male -0.1506 0.02649 -5.685 -0.2025 -0.09867 fixed
sibling_count3 0.01379 0.04149 0.3323 -0.06754 0.09511 fixed
sibling_count4 -0.06312 0.04509 -1.4 -0.1515 0.02524 fixed
sibling_count5 -0.1219 0.05343 -2.282 -0.2266 -0.01718 fixed
sibling_count5+ -0.2157 0.04628 -4.662 -0.3064 -0.125 fixed
sd_(Intercept).mother_pidlink 0.4422 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9234 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5462 0.4544 -1.202 -1.437 0.3443 fixed
birth_order 0.01591 0.008989 1.77 -0.001704 0.03353 fixed
poly(age, 3, raw = TRUE)1 0.1127 0.05159 2.185 0.01163 0.2138 fixed
poly(age, 3, raw = TRUE)2 -0.004158 0.001842 -2.258 -0.007768 -0.0005481 fixed
poly(age, 3, raw = TRUE)3 0.0000456 0.00002088 2.184 0.000004674 0.00008652 fixed
male -0.151 0.02648 -5.703 -0.2029 -0.09913 fixed
sibling_count3 0.005873 0.04173 0.1408 -0.07591 0.08765 fixed
sibling_count4 -0.08142 0.04625 -1.76 -0.1721 0.009226 fixed
sibling_count5 -0.1509 0.05588 -2.701 -0.2604 -0.04139 fixed
sibling_count5+ -0.2759 0.0574 -4.806 -0.3884 -0.1634 fixed
sd_(Intercept).mother_pidlink 0.4423 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9231 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5386 0.4552 -1.183 -1.431 0.3536 fixed
poly(age, 3, raw = TRUE)1 0.1138 0.05163 2.205 0.01263 0.215 fixed
poly(age, 3, raw = TRUE)2 -0.004199 0.001843 -2.278 -0.007812 -0.0005865 fixed
poly(age, 3, raw = TRUE)3 0.00004613 0.0000209 2.207 0.000005169 0.00008709 fixed
male -0.1512 0.02649 -5.709 -0.2032 -0.09932 fixed
sibling_count3 -0.002021 0.04257 -0.04749 -0.08546 0.08141 fixed
sibling_count4 -0.09644 0.04802 -2.008 -0.1906 -0.002312 fixed
sibling_count5 -0.1759 0.0583 -3.017 -0.2901 -0.06163 fixed
sibling_count5+ -0.2909 0.05907 -4.924 -0.4067 -0.1751 fixed
birth_order_nonlinear2 0.01337 0.03353 0.3987 -0.05235 0.07909 fixed
birth_order_nonlinear3 0.06506 0.04154 1.566 -0.01636 0.1465 fixed
birth_order_nonlinear4 0.08177 0.05293 1.545 -0.02197 0.1855 fixed
birth_order_nonlinear5 0.1201 0.06705 1.791 -0.01132 0.2515 fixed
birth_order_nonlinear5+ 0.09651 0.06699 1.441 -0.03479 0.2278 fixed
sd_(Intercept).mother_pidlink 0.4415 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9236 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5806 0.4565 -1.272 -1.475 0.3142 fixed
poly(age, 3, raw = TRUE)1 0.1187 0.05177 2.292 0.0172 0.2201 fixed
poly(age, 3, raw = TRUE)2 -0.004378 0.001849 -2.368 -0.008002 -0.0007549 fixed
poly(age, 3, raw = TRUE)3 0.00004819 0.00002096 2.299 0.000007106 0.00008928 fixed
male -0.15 0.02652 -5.658 -0.202 -0.09806 fixed
count_birth_order2/2 0.01536 0.05925 0.2592 -0.1008 0.1315 fixed
count_birth_order1/3 0.003387 0.05402 0.0627 -0.1025 0.1093 fixed
count_birth_order2/3 0.008751 0.05945 0.1472 -0.1078 0.1253 fixed
count_birth_order3/3 0.05884 0.06529 0.9012 -0.06912 0.1868 fixed
count_birth_order1/4 -0.09925 0.06679 -1.486 -0.2302 0.03165 fixed
count_birth_order2/4 -0.123 0.06896 -1.783 -0.2581 0.01218 fixed
count_birth_order3/4 0.001164 0.07194 0.01618 -0.1398 0.1422 fixed
count_birth_order4/4 0.008341 0.07586 0.11 -0.1403 0.157 fixed
count_birth_order1/5 -0.123 0.09033 -1.361 -0.3 0.05408 fixed
count_birth_order2/5 -0.0875 0.0999 -0.8758 -0.2833 0.1083 fixed
count_birth_order3/5 -0.0931 0.09501 -0.9799 -0.2793 0.09312 fixed
count_birth_order4/5 -0.1781 0.09211 -1.934 -0.3587 0.002391 fixed
count_birth_order5/5 -0.11 0.09724 -1.131 -0.3006 0.08061 fixed
count_birth_order1/5+ -0.3524 0.09201 -3.83 -0.5328 -0.1721 fixed
count_birth_order2/5+ -0.2463 0.09116 -2.702 -0.425 -0.06767 fixed
count_birth_order3/5+ -0.2817 0.09003 -3.129 -0.4582 -0.1053 fixed
count_birth_order4/5+ -0.1651 0.08801 -1.876 -0.3376 0.007367 fixed
count_birth_order5/5+ -0.1366 0.08107 -1.685 -0.2955 0.02229 fixed
count_birth_order5+/5+ -0.1927 0.06138 -3.139 -0.313 -0.07237 fixed
sd_(Intercept).mother_pidlink 0.4406 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9244 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 16690 16763 -8334 16668 NA NA NA
12 16689 16769 -8332 16665 3.138 1 0.07648
16 16695 16801 -8331 16663 2.443 4 0.6549
26 16709 16882 -8329 16657 5.429 10 0.8607

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Raven 2015 young

birthorder <- birthorder %>% mutate(outcome = raven_2015_young)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.913 0.1161 -25.09 -3.141 -2.686 fixed
poly(age, 3, raw = TRUE)1 0.4204 0.02155 19.5 0.3781 0.4626 fixed
poly(age, 3, raw = TRUE)2 -0.01677 0.00125 -13.41 -0.01922 -0.01432 fixed
poly(age, 3, raw = TRUE)3 0.000206 0.00002297 8.97 0.000161 0.0002511 fixed
male 0.04925 0.01654 2.977 0.01683 0.08168 fixed
sibling_count3 -0.03514 0.02866 -1.226 -0.09131 0.02103 fixed
sibling_count4 -0.08887 0.03214 -2.765 -0.1519 -0.02588 fixed
sibling_count5 -0.07689 0.03636 -2.115 -0.1482 -0.005637 fixed
sibling_count5+ -0.244 0.02882 -8.467 -0.3005 -0.1875 fixed
sd_(Intercept).mother_pidlink 0.5099 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.786 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.899 0.1165 -24.88 -3.127 -2.671 fixed
birth_order -0.006508 0.004435 -1.467 -0.0152 0.002184 fixed
poly(age, 3, raw = TRUE)1 0.4199 0.02156 19.48 0.3777 0.4622 fixed
poly(age, 3, raw = TRUE)2 -0.01677 0.00125 -13.41 -0.01922 -0.01432 fixed
poly(age, 3, raw = TRUE)3 0.0002058 0.00002297 8.96 0.0001608 0.0002508 fixed
male 0.04936 0.01654 2.984 0.01693 0.08178 fixed
sibling_count3 -0.03129 0.02878 -1.087 -0.08769 0.02511 fixed
sibling_count4 -0.07961 0.03275 -2.431 -0.1438 -0.01543 fixed
sibling_count5 -0.06196 0.03775 -1.641 -0.1359 0.01202 fixed
sibling_count5+ -0.2073 0.03816 -5.432 -0.2821 -0.1325 fixed
sd_(Intercept).mother_pidlink 0.5097 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7861 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.907 0.1177 -24.7 -3.137 -2.676 fixed
poly(age, 3, raw = TRUE)1 0.4201 0.02164 19.42 0.3777 0.4625 fixed
poly(age, 3, raw = TRUE)2 -0.01679 0.001252 -13.41 -0.01924 -0.01433 fixed
poly(age, 3, raw = TRUE)3 0.0002063 0.00002299 8.976 0.0001613 0.0002514 fixed
male 0.04912 0.01655 2.969 0.01669 0.08155 fixed
sibling_count3 -0.03626 0.02964 -1.223 -0.09435 0.02184 fixed
sibling_count4 -0.08243 0.03527 -2.337 -0.1515 -0.01331 fixed
sibling_count5 -0.05136 0.04155 -1.236 -0.1328 0.03007 fixed
sibling_count5+ -0.1992 0.04201 -4.74 -0.2815 -0.1168 fixed
birth_order_nonlinear2 -0.0004814 0.02252 -0.02137 -0.04463 0.04367 fixed
birth_order_nonlinear3 0.007394 0.02843 0.2601 -0.04832 0.06311 fixed
birth_order_nonlinear4 -0.02942 0.03523 -0.835 -0.09847 0.03963 fixed
birth_order_nonlinear5 -0.06717 0.04142 -1.622 -0.1484 0.01401 fixed
birth_order_nonlinear5+ -0.05048 0.03943 -1.28 -0.1278 0.02681 fixed
sd_(Intercept).mother_pidlink 0.51 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.786 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.913 0.1186 -24.56 -3.146 -2.681 fixed
poly(age, 3, raw = TRUE)1 0.4217 0.02167 19.46 0.3792 0.4642 fixed
poly(age, 3, raw = TRUE)2 -0.01689 0.001254 -13.47 -0.01934 -0.01443 fixed
poly(age, 3, raw = TRUE)3 0.000208 0.00002301 9.043 0.000163 0.0002531 fixed
male 0.04874 0.01655 2.945 0.01631 0.08117 fixed
count_birth_order2/2 -0.004616 0.0367 -0.1258 -0.07655 0.06732 fixed
count_birth_order1/3 -0.06756 0.03753 -1.8 -0.1411 0.005986 fixed
count_birth_order2/3 -0.03873 0.03826 -1.012 -0.1137 0.03627 fixed
count_birth_order3/3 0.02046 0.04332 0.4722 -0.06446 0.1054 fixed
count_birth_order1/4 -0.06759 0.05129 -1.318 -0.1681 0.03294 fixed
count_birth_order2/4 -0.04354 0.04874 -0.8935 -0.1391 0.05198 fixed
count_birth_order3/4 -0.1544 0.04591 -3.362 -0.2444 -0.06439 fixed
count_birth_order4/4 -0.07384 0.05012 -1.473 -0.1721 0.02438 fixed
count_birth_order1/5 0.0104 0.0696 0.1495 -0.126 0.1468 fixed
count_birth_order2/5 -0.08665 0.06625 -1.308 -0.2165 0.0432 fixed
count_birth_order3/5 -0.01868 0.06002 -0.3112 -0.1363 0.09896 fixed
count_birth_order4/5 -0.1326 0.05643 -2.35 -0.2432 -0.02201 fixed
count_birth_order5/5 -0.1062 0.05597 -1.897 -0.2159 0.003508 fixed
count_birth_order1/5+ -0.1548 0.0697 -2.221 -0.2914 -0.01816 fixed
count_birth_order2/5+ -0.2376 0.06527 -3.64 -0.3655 -0.1096 fixed
count_birth_order3/5+ -0.1763 0.0581 -3.035 -0.2902 -0.06244 fixed
count_birth_order4/5+ -0.2312 0.05287 -4.372 -0.3348 -0.1275 fixed
count_birth_order5/5+ -0.2756 0.0475 -5.802 -0.3687 -0.1825 fixed
count_birth_order5+/5+ -0.2508 0.03316 -7.564 -0.3158 -0.1858 fixed
sd_(Intercept).mother_pidlink 0.5098 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7859 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 31548 31629 -15763 31526 NA NA NA
12 31548 31636 -15762 31524 2.156 1 0.142
16 31554 31672 -15761 31522 2.168 4 0.705
26 31560 31752 -15754 31508 14.05 10 0.1707

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.135 0.1776 -17.65 -3.483 -2.787 fixed
poly(age, 3, raw = TRUE)1 0.4673 0.03646 12.82 0.3958 0.5388 fixed
poly(age, 3, raw = TRUE)2 -0.01963 0.00231 -8.494 -0.02415 -0.0151 fixed
poly(age, 3, raw = TRUE)3 0.0002649 0.0000455 5.821 0.0001757 0.0003541 fixed
male 0.03061 0.01923 1.592 -0.007084 0.06831 fixed
sibling_count3 -0.05946 0.02726 -2.182 -0.1129 -0.00604 fixed
sibling_count4 -0.1194 0.03344 -3.571 -0.185 -0.05388 fixed
sibling_count5 -0.2169 0.04292 -5.054 -0.301 -0.1328 fixed
sibling_count5+ -0.2597 0.03991 -6.508 -0.338 -0.1815 fixed
sd_(Intercept).mother_pidlink 0.4912 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7717 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.103 0.1799 -17.25 -3.456 -2.751 fixed
birth_order -0.009805 0.008629 -1.136 -0.02672 0.007108 fixed
poly(age, 3, raw = TRUE)1 0.4643 0.03656 12.7 0.3927 0.536 fixed
poly(age, 3, raw = TRUE)2 -0.01948 0.002314 -8.417 -0.02401 -0.01494 fixed
poly(age, 3, raw = TRUE)3 0.0002623 0.00004557 5.756 0.000173 0.0003516 fixed
male 0.0307 0.01923 1.596 -0.006995 0.0684 fixed
sibling_count3 -0.05286 0.02786 -1.897 -0.1075 0.001741 fixed
sibling_count4 -0.1048 0.03583 -2.924 -0.175 -0.03455 fixed
sibling_count5 -0.1926 0.04793 -4.018 -0.2865 -0.09866 fixed
sibling_count5+ -0.2124 0.05769 -3.682 -0.3255 -0.09936 fixed
sd_(Intercept).mother_pidlink 0.4907 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7719 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.137 0.1795 -17.48 -3.488 -2.785 fixed
poly(age, 3, raw = TRUE)1 0.4676 0.03657 12.79 0.3959 0.5393 fixed
poly(age, 3, raw = TRUE)2 -0.01965 0.002314 -8.493 -0.02419 -0.01512 fixed
poly(age, 3, raw = TRUE)3 0.0002655 0.00004556 5.826 0.0001761 0.0003548 fixed
male 0.02997 0.01924 1.558 -0.007733 0.06767 fixed
sibling_count3 -0.0652 0.02908 -2.242 -0.1222 -0.008209 fixed
sibling_count4 -0.1234 0.03863 -3.194 -0.1991 -0.04767 fixed
sibling_count5 -0.1819 0.05257 -3.459 -0.2849 -0.07882 fixed
sibling_count5+ -0.2346 0.06227 -3.768 -0.3567 -0.1126 fixed
birth_order_nonlinear2 0.0033 0.02304 0.1432 -0.04186 0.04846 fixed
birth_order_nonlinear3 0.0203 0.03226 0.6294 -0.04292 0.08353 fixed
birth_order_nonlinear4 -0.002993 0.04418 -0.06775 -0.08958 0.08359 fixed
birth_order_nonlinear5 -0.1088 0.05943 -1.831 -0.2253 0.00765 fixed
birth_order_nonlinear5+ -0.01069 0.06881 -0.1554 -0.1456 0.1242 fixed
sd_(Intercept).mother_pidlink 0.4915 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7716 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.146 0.1802 -17.46 -3.499 -2.793 fixed
poly(age, 3, raw = TRUE)1 0.4697 0.03664 12.82 0.3979 0.5415 fixed
poly(age, 3, raw = TRUE)2 -0.01981 0.002319 -8.545 -0.02436 -0.01527 fixed
poly(age, 3, raw = TRUE)3 0.000269 0.00004566 5.891 0.0001795 0.0003585 fixed
male 0.03042 0.01924 1.581 -0.007291 0.06813 fixed
count_birth_order2/2 0.006774 0.03232 0.2096 -0.05657 0.07011 fixed
count_birth_order1/3 -0.08009 0.038 -2.108 -0.1546 -0.00561 fixed
count_birth_order2/3 -0.07185 0.03655 -1.966 -0.1435 -0.0002115 fixed
count_birth_order3/3 -0.01001 0.04001 -0.2501 -0.08843 0.06842 fixed
count_birth_order1/4 -0.04987 0.05992 -0.8323 -0.1673 0.06757 fixed
count_birth_order2/4 -0.09864 0.05259 -1.876 -0.2017 0.00444 fixed
count_birth_order3/4 -0.1871 0.05021 -3.726 -0.2855 -0.08867 fixed
count_birth_order4/4 -0.1056 0.04973 -2.124 -0.2031 -0.008146 fixed
count_birth_order1/5 -0.2007 0.1043 -1.924 -0.4053 0.003786 fixed
count_birth_order2/5 -0.146 0.09299 -1.57 -0.3283 0.03624 fixed
count_birth_order3/5 -0.08258 0.07873 -1.049 -0.2369 0.07173 fixed
count_birth_order4/5 -0.2497 0.06518 -3.83 -0.3774 -0.1219 fixed
count_birth_order5/5 -0.2776 0.06325 -4.388 -0.4016 -0.1536 fixed
count_birth_order1/5+ -0.268 0.1333 -2.01 -0.5293 -0.006699 fixed
count_birth_order2/5+ -0.262 0.122 -2.148 -0.5011 -0.02298 fixed
count_birth_order3/5+ -0.2095 0.1009 -2.077 -0.4072 -0.0118 fixed
count_birth_order4/5+ -0.179 0.09054 -1.977 -0.3564 -0.001516 fixed
count_birth_order5/5+ -0.359 0.07278 -4.933 -0.5016 -0.2164 fixed
count_birth_order5+/5+ -0.244 0.04847 -5.033 -0.339 -0.149 fixed
sd_(Intercept).mother_pidlink 0.4921 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7711 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 22246 22323 -11112 22224 NA NA NA
12 22246 22331 -11111 22222 1.293 1 0.2554
16 22250 22363 -11109 22218 4.165 4 0.3842
26 22259 22442 -11103 22207 11.6 10 0.3129

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.141 0.177 -17.75 -3.488 -2.794 fixed
poly(age, 3, raw = TRUE)1 0.4677 0.03633 12.87 0.3965 0.5389 fixed
poly(age, 3, raw = TRUE)2 -0.01973 0.002303 -8.564 -0.02424 -0.01521 fixed
poly(age, 3, raw = TRUE)3 0.0002666 0.00004538 5.876 0.0001777 0.0003556 fixed
male 0.02784 0.01915 1.454 -0.009698 0.06538 fixed
sibling_count3 -0.04184 0.0286 -1.463 -0.09789 0.0142 fixed
sibling_count4 -0.07294 0.03364 -2.168 -0.1389 -0.007009 fixed
sibling_count5 -0.08591 0.03944 -2.178 -0.1632 -0.008609 fixed
sibling_count5+ -0.1944 0.03514 -5.532 -0.2633 -0.1255 fixed
sd_(Intercept).mother_pidlink 0.4946 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7717 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.082 0.1787 -17.25 -3.432 -2.732 fixed
birth_order -0.0174 0.007239 -2.404 -0.03159 -0.003213 fixed
poly(age, 3, raw = TRUE)1 0.4622 0.0364 12.7 0.3908 0.5335 fixed
poly(age, 3, raw = TRUE)2 -0.01947 0.002306 -8.443 -0.02399 -0.01495 fixed
poly(age, 3, raw = TRUE)3 0.0002621 0.00004541 5.772 0.0001731 0.0003511 fixed
male 0.02841 0.01915 1.484 -0.009121 0.06595 fixed
sibling_count3 -0.03073 0.02895 -1.062 -0.08747 0.02601 fixed
sibling_count4 -0.04801 0.03518 -1.364 -0.117 0.02095 fixed
sibling_count5 -0.04672 0.04264 -1.096 -0.1303 0.03686 fixed
sibling_count5+ -0.1149 0.04825 -2.381 -0.2095 -0.02034 fixed
sd_(Intercept).mother_pidlink 0.4935 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.772 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.112 0.1786 -17.42 -3.462 -2.762 fixed
poly(age, 3, raw = TRUE)1 0.4643 0.03643 12.75 0.3929 0.5357 fixed
poly(age, 3, raw = TRUE)2 -0.01959 0.002306 -8.495 -0.02411 -0.01507 fixed
poly(age, 3, raw = TRUE)3 0.0002645 0.00004542 5.823 0.0001754 0.0003535 fixed
male 0.0276 0.01915 1.441 -0.009942 0.06514 fixed
sibling_count3 -0.03796 0.0301 -1.261 -0.09696 0.02103 fixed
sibling_count4 -0.05896 0.03796 -1.553 -0.1334 0.01543 fixed
sibling_count5 -0.03642 0.0465 -0.7831 -0.1276 0.05473 fixed
sibling_count5+ -0.1248 0.05121 -2.436 -0.2251 -0.02438 fixed
birth_order_nonlinear2 -0.01198 0.02373 -0.5048 -0.05848 0.03453 fixed
birth_order_nonlinear3 -0.00978 0.03163 -0.3092 -0.07177 0.05221 fixed
birth_order_nonlinear4 -0.03695 0.04151 -0.8903 -0.1183 0.0444 fixed
birth_order_nonlinear5 -0.132 0.05216 -2.532 -0.2343 -0.02981 fixed
birth_order_nonlinear5+ -0.08421 0.05662 -1.487 -0.1952 0.02676 fixed
sd_(Intercept).mother_pidlink 0.4944 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7717 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.121 0.1796 -17.38 -3.473 -2.769 fixed
poly(age, 3, raw = TRUE)1 0.467 0.03649 12.8 0.3955 0.5385 fixed
poly(age, 3, raw = TRUE)2 -0.0198 0.00231 -8.57 -0.02433 -0.01527 fixed
poly(age, 3, raw = TRUE)3 0.000269 0.00004549 5.914 0.0001799 0.0003582 fixed
male 0.02854 0.01916 1.489 -0.009018 0.0661 fixed
count_birth_order2/2 -0.01543 0.03533 -0.4366 -0.08468 0.05382 fixed
count_birth_order1/3 -0.06278 0.03938 -1.594 -0.14 0.0144 fixed
count_birth_order2/3 -0.05192 0.03852 -1.348 -0.1274 0.02357 fixed
count_birth_order3/3 -0.0183 0.04222 -0.4335 -0.101 0.06445 fixed
count_birth_order1/4 -0.05623 0.05864 -0.9588 -0.1712 0.05871 fixed
count_birth_order2/4 -0.0529 0.05265 -1.005 -0.1561 0.0503 fixed
count_birth_order3/4 -0.1333 0.0504 -2.644 -0.2321 -0.03448 fixed
count_birth_order4/4 -0.05557 0.05097 -1.09 -0.1555 0.04433 fixed
count_birth_order1/5 0.07905 0.08526 0.9271 -0.08807 0.2462 fixed
count_birth_order2/5 -0.03706 0.07632 -0.4856 -0.1866 0.1125 fixed
count_birth_order3/5 -0.03526 0.06854 -0.5144 -0.1696 0.09908 fixed
count_birth_order4/5 -0.1447 0.06314 -2.291 -0.2684 -0.02093 fixed
count_birth_order5/5 -0.1711 0.06082 -2.813 -0.2903 -0.05186 fixed
count_birth_order1/5+ -0.1041 0.09126 -1.14 -0.2829 0.0748 fixed
count_birth_order2/5+ -0.2198 0.1014 -2.167 -0.4185 -0.02104 fixed
count_birth_order3/5+ -0.1055 0.07852 -1.344 -0.2594 0.04834 fixed
count_birth_order4/5+ -0.1646 0.07207 -2.283 -0.3058 -0.02331 fixed
count_birth_order5/5+ -0.2587 0.06275 -4.123 -0.3817 -0.1357 fixed
count_birth_order5+/5+ -0.2101 0.04393 -4.782 -0.2962 -0.124 fixed
sd_(Intercept).mother_pidlink 0.4947 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7716 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 22545 22623 -11262 22523 NA NA NA
12 22542 22626 -11259 22518 5.783 1 0.01618
16 22548 22661 -11258 22516 1.276 4 0.8654
26 22558 22742 -11253 22506 9.852 10 0.4535

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.112 0.1791 -17.38 -3.463 -2.761 fixed
poly(age, 3, raw = TRUE)1 0.4625 0.03675 12.58 0.3904 0.5345 fixed
poly(age, 3, raw = TRUE)2 -0.0194 0.002327 -8.337 -0.02397 -0.01484 fixed
poly(age, 3, raw = TRUE)3 0.0002614 0.00004581 5.706 0.0001716 0.0003512 fixed
male 0.03031 0.01944 1.559 -0.007785 0.0684 fixed
sibling_count3 -0.04685 0.02732 -1.715 -0.1004 0.006702 fixed
sibling_count4 -0.09949 0.0338 -2.944 -0.1657 -0.03324 fixed
sibling_count5 -0.2019 0.04473 -4.514 -0.2896 -0.1143 fixed
sibling_count5+ -0.2519 0.04129 -6.1 -0.3328 -0.171 fixed
sd_(Intercept).mother_pidlink 0.4914 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7689 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.086 0.1816 -17 -3.442 -2.73 fixed
birth_order -0.00767 0.008849 -0.8668 -0.02501 0.009673 fixed
poly(age, 3, raw = TRUE)1 0.46 0.03687 12.48 0.3877 0.5322 fixed
poly(age, 3, raw = TRUE)2 -0.01928 0.002332 -8.267 -0.02385 -0.01471 fixed
poly(age, 3, raw = TRUE)3 0.0002592 0.00004589 5.648 0.0001692 0.0003491 fixed
male 0.03038 0.01944 1.563 -0.007719 0.06848 fixed
sibling_count3 -0.04173 0.02795 -1.493 -0.09651 0.01304 fixed
sibling_count4 -0.08812 0.03625 -2.431 -0.1592 -0.01706 fixed
sibling_count5 -0.1833 0.04961 -3.695 -0.2805 -0.08609 fixed
sibling_count5+ -0.2153 0.05902 -3.649 -0.331 -0.09967 fixed
sd_(Intercept).mother_pidlink 0.4911 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7691 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.115 0.181 -17.21 -3.47 -2.76 fixed
poly(age, 3, raw = TRUE)1 0.4631 0.03688 12.56 0.3908 0.5354 fixed
poly(age, 3, raw = TRUE)2 -0.01947 0.002332 -8.347 -0.02404 -0.0149 fixed
poly(age, 3, raw = TRUE)3 0.0002629 0.00004589 5.729 0.000173 0.0003529 fixed
male 0.03004 0.01944 1.545 -0.008058 0.06814 fixed
sibling_count3 -0.04791 0.02914 -1.644 -0.105 0.009209 fixed
sibling_count4 -0.102 0.03901 -2.615 -0.1785 -0.02556 fixed
sibling_count5 -0.1672 0.05419 -3.085 -0.2734 -0.06097 fixed
sibling_count5+ -0.2479 0.06378 -3.887 -0.3729 -0.1229 fixed
birth_order_nonlinear2 0.003997 0.02302 0.1737 -0.04112 0.04911 fixed
birth_order_nonlinear3 0.003994 0.03243 0.1232 -0.05957 0.06756 fixed
birth_order_nonlinear4 0.006891 0.04507 0.1529 -0.08144 0.09523 fixed
birth_order_nonlinear5 -0.1106 0.06204 -1.782 -0.2322 0.01102 fixed
birth_order_nonlinear5+ 0.02731 0.07101 0.3846 -0.1119 0.1665 fixed
sd_(Intercept).mother_pidlink 0.4917 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7688 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.121 0.1818 -17.17 -3.478 -2.765 fixed
poly(age, 3, raw = TRUE)1 0.4648 0.03695 12.58 0.3924 0.5372 fixed
poly(age, 3, raw = TRUE)2 -0.01959 0.002337 -8.384 -0.02417 -0.01501 fixed
poly(age, 3, raw = TRUE)3 0.0002656 0.00004598 5.776 0.0001754 0.0003557 fixed
male 0.03075 0.01944 1.582 -0.00735 0.06886 fixed
count_birth_order2/2 0.001768 0.03197 0.0553 -0.06088 0.06442 fixed
count_birth_order1/3 -0.06413 0.03815 -1.681 -0.1389 0.01064 fixed
count_birth_order2/3 -0.05242 0.03664 -1.431 -0.1242 0.0194 fixed
count_birth_order3/3 -0.0167 0.04017 -0.4156 -0.09543 0.06203 fixed
count_birth_order1/4 -0.0457 0.06025 -0.7585 -0.1638 0.07239 fixed
count_birth_order2/4 -0.07186 0.05325 -1.349 -0.1762 0.03251 fixed
count_birth_order3/4 -0.1843 0.05056 -3.645 -0.2834 -0.08519 fixed
count_birth_order4/4 -0.07292 0.05041 -1.447 -0.1717 0.02587 fixed
count_birth_order1/5 -0.1909 0.1042 -1.832 -0.3952 0.01334 fixed
count_birth_order2/5 -0.1379 0.09972 -1.383 -0.3334 0.05753 fixed
count_birth_order3/5 -0.07514 0.0808 -0.9299 -0.2335 0.08323 fixed
count_birth_order4/5 -0.2494 0.06853 -3.639 -0.3837 -0.115 fixed
count_birth_order5/5 -0.2528 0.06699 -3.774 -0.3841 -0.1215 fixed
count_birth_order1/5+ -0.2926 0.1343 -2.179 -0.5557 -0.02945 fixed
count_birth_order2/5+ -0.2989 0.1239 -2.412 -0.5418 -0.05605 fixed
count_birth_order3/5+ -0.2104 0.1014 -2.076 -0.4091 -0.01174 fixed
count_birth_order4/5+ -0.1581 0.09617 -1.644 -0.3466 0.03035 fixed
count_birth_order5/5+ -0.3945 0.07643 -5.161 -0.5443 -0.2447 fixed
count_birth_order5+/5+ -0.2215 0.05033 -4.401 -0.3202 -0.1229 fixed
sd_(Intercept).mother_pidlink 0.4926 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7681 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 21616 21693 -10797 21594 NA NA NA
12 21617 21701 -10796 21593 0.753 1 0.3855
16 21620 21732 -10794 21588 4.805 4 0.3079
26 21627 21810 -10788 21575 12.77 10 0.2369

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Numeracy 2015 young

birthorder <- birthorder %>% mutate(outcome = math_2015_young)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.466 0.1215 -20.3 -2.704 -2.228 fixed
poly(age, 3, raw = TRUE)1 0.4048 0.02252 17.97 0.3606 0.4489 fixed
poly(age, 3, raw = TRUE)2 -0.01815 0.001303 -13.93 -0.02071 -0.0156 fixed
poly(age, 3, raw = TRUE)3 0.0002515 0.00002388 10.53 0.0002047 0.0002983 fixed
male -0.1617 0.01737 -9.306 -0.1957 -0.1276 fixed
sibling_count3 -0.003588 0.02894 -0.124 -0.06031 0.05313 fixed
sibling_count4 -0.08819 0.03232 -2.729 -0.1515 -0.02485 fixed
sibling_count5 -0.04881 0.0364 -1.341 -0.1202 0.02254 fixed
sibling_count5+ -0.2076 0.02899 -7.16 -0.2644 -0.1508 fixed
sd_(Intercept).mother_pidlink 0.4412 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.857 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.454 0.1219 -20.14 -2.693 -2.215 fixed
birth_order -0.00565 0.004521 -1.25 -0.01451 0.003212 fixed
poly(age, 3, raw = TRUE)1 0.4044 0.02252 17.95 0.3602 0.4485 fixed
poly(age, 3, raw = TRUE)2 -0.01814 0.001303 -13.93 -0.0207 -0.01559 fixed
poly(age, 3, raw = TRUE)3 0.0002512 0.00002388 10.52 0.0002044 0.000298 fixed
male -0.1616 0.01737 -9.302 -0.1956 -0.1275 fixed
sibling_count3 -0.000338 0.02905 -0.01163 -0.05728 0.05661 fixed
sibling_count4 -0.08042 0.0329 -2.444 -0.1449 -0.01593 fixed
sibling_count5 -0.03629 0.03775 -0.9613 -0.1103 0.0377 fixed
sibling_count5+ -0.1764 0.03825 -4.612 -0.2514 -0.1014 fixed
sd_(Intercept).mother_pidlink 0.4408 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8571 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.481 0.1229 -20.18 -2.722 -2.24 fixed
poly(age, 3, raw = TRUE)1 0.4066 0.02259 18 0.3624 0.4509 fixed
poly(age, 3, raw = TRUE)2 -0.01824 0.001304 -13.98 -0.02079 -0.01568 fixed
poly(age, 3, raw = TRUE)3 0.0002523 0.00002389 10.56 0.0002055 0.0002992 fixed
male -0.1618 0.01737 -9.315 -0.1959 -0.1278 fixed
sibling_count3 -0.01878 0.03001 -0.6259 -0.07759 0.04003 fixed
sibling_count4 -0.09841 0.03565 -2.76 -0.1683 -0.02853 fixed
sibling_count5 -0.04989 0.04188 -1.191 -0.132 0.03219 fixed
sibling_count5+ -0.1826 0.04255 -4.29 -0.266 -0.09916 fixed
birth_order_nonlinear2 0.01111 0.02393 0.4642 -0.03579 0.058 fixed
birth_order_nonlinear3 0.06162 0.03008 2.048 0.002657 0.1206 fixed
birth_order_nonlinear4 -0.02853 0.03724 -0.7662 -0.1015 0.04445 fixed
birth_order_nonlinear5 -0.008226 0.04369 -0.1883 -0.09387 0.07741 fixed
birth_order_nonlinear5+ -0.03737 0.04097 -0.912 -0.1177 0.04294 fixed
sd_(Intercept).mother_pidlink 0.4416 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8566 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.495 0.1239 -20.13 -2.738 -2.252 fixed
poly(age, 3, raw = TRUE)1 0.4073 0.02263 18 0.363 0.4517 fixed
poly(age, 3, raw = TRUE)2 -0.01828 0.001306 -13.99 -0.02084 -0.01572 fixed
poly(age, 3, raw = TRUE)3 0.0002531 0.00002392 10.58 0.0002062 0.0003 fixed
male -0.1617 0.01738 -9.306 -0.1958 -0.1277 fixed
count_birth_order2/2 0.0442 0.0389 1.136 -0.03205 0.1204 fixed
count_birth_order1/3 0.003874 0.03882 0.09977 -0.07222 0.07997 fixed
count_birth_order2/3 0.01011 0.03964 0.2552 -0.06758 0.08781 fixed
count_birth_order3/3 0.02486 0.04499 0.5525 -0.06332 0.113 fixed
count_birth_order1/4 -0.09987 0.05334 -1.872 -0.2044 0.004673 fixed
count_birth_order2/4 -0.08891 0.05068 -1.754 -0.1882 0.01042 fixed
count_birth_order3/4 -0.031 0.04767 -0.6502 -0.1244 0.06244 fixed
count_birth_order4/4 -0.0841 0.05214 -1.613 -0.1863 0.01808 fixed
count_birth_order1/5 -0.03411 0.07281 -0.4685 -0.1768 0.1086 fixed
count_birth_order2/5 -0.04574 0.06923 -0.6606 -0.1814 0.08996 fixed
count_birth_order3/5 0.04814 0.06259 0.7691 -0.07453 0.1708 fixed
count_birth_order4/5 -0.1116 0.05873 -1.9 -0.2267 0.003508 fixed
count_birth_order5/5 -0.01722 0.05821 -0.2958 -0.1313 0.09687 fixed
count_birth_order1/5+ -0.1167 0.07317 -1.595 -0.2601 0.02668 fixed
count_birth_order2/5+ -0.2274 0.06847 -3.322 -0.3616 -0.09323 fixed
count_birth_order3/5+ -0.06418 0.06079 -1.056 -0.1833 0.05497 fixed
count_birth_order4/5+ -0.1983 0.05525 -3.589 -0.3066 -0.09001 fixed
count_birth_order5/5+ -0.1984 0.04952 -4.007 -0.2954 -0.1013 fixed
count_birth_order5+/5+ -0.209 0.03399 -6.149 -0.2756 -0.1424 fixed
sd_(Intercept).mother_pidlink 0.4408 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8571 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 32480 32561 -16229 32458 NA NA NA
12 32480 32569 -16228 32456 1.564 1 0.2111
16 32480 32598 -16224 32448 7.826 4 0.09815
26 32494 32686 -16221 32442 6.537 10 0.7683

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.104 0.1915 -16.21 -3.48 -2.729 fixed
poly(age, 3, raw = TRUE)1 0.5371 0.03932 13.66 0.4601 0.6142 fixed
poly(age, 3, raw = TRUE)2 -0.02639 0.002491 -10.59 -0.03128 -0.02151 fixed
poly(age, 3, raw = TRUE)3 0.0004132 0.00004907 8.422 0.0003171 0.0005094 fixed
male -0.1782 0.0206 -8.647 -0.2185 -0.1378 fixed
sibling_count3 -0.004614 0.02826 -0.1633 -0.06001 0.05078 fixed
sibling_count4 -0.0559 0.03451 -1.62 -0.1235 0.01174 fixed
sibling_count5 -0.1578 0.04412 -3.575 -0.2442 -0.07127 fixed
sibling_count5+ -0.2106 0.04109 -5.126 -0.2911 -0.1301 fixed
sd_(Intercept).mother_pidlink 0.4467 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8555 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.11 0.1937 -16.06 -3.49 -2.731 fixed
birth_order 0.001986 0.009122 0.2177 -0.01589 0.01987 fixed
poly(age, 3, raw = TRUE)1 0.5377 0.03941 13.64 0.4605 0.615 fixed
poly(age, 3, raw = TRUE)2 -0.02642 0.002495 -10.59 -0.03131 -0.02153 fixed
poly(age, 3, raw = TRUE)3 0.0004138 0.00004912 8.423 0.0003175 0.00051 fixed
male -0.1782 0.0206 -8.647 -0.2186 -0.1378 fixed
sibling_count3 -0.005917 0.02889 -0.2048 -0.06254 0.05071 fixed
sibling_count4 -0.05879 0.03699 -1.589 -0.1313 0.0137 fixed
sibling_count5 -0.1626 0.04936 -3.293 -0.2593 -0.06582 fixed
sibling_count5+ -0.22 0.05974 -3.683 -0.3371 -0.1029 fixed
sd_(Intercept).mother_pidlink 0.4468 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8555 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.126 0.1933 -16.18 -3.505 -2.747 fixed
poly(age, 3, raw = TRUE)1 0.5394 0.03943 13.68 0.4621 0.6166 fixed
poly(age, 3, raw = TRUE)2 -0.02653 0.002495 -10.63 -0.03142 -0.02164 fixed
poly(age, 3, raw = TRUE)3 0.0004159 0.00004913 8.465 0.0003196 0.0005122 fixed
male -0.1786 0.02061 -8.664 -0.2189 -0.1382 fixed
sibling_count3 -0.00646 0.03024 -0.2136 -0.06573 0.05281 fixed
sibling_count4 -0.05353 0.04007 -1.336 -0.1321 0.025 fixed
sibling_count5 -0.1354 0.05447 -2.485 -0.2421 -0.02859 fixed
sibling_count5+ -0.2334 0.06479 -3.603 -0.3604 -0.1065 fixed
birth_order_nonlinear2 0.02891 0.02498 1.157 -0.02005 0.07787 fixed
birth_order_nonlinear3 0.004704 0.03478 0.1352 -0.06347 0.07288 fixed
birth_order_nonlinear4 -0.002522 0.04752 -0.05308 -0.09565 0.09061 fixed
birth_order_nonlinear5 -0.04888 0.06391 -0.7648 -0.1741 0.07638 fixed
birth_order_nonlinear5+ 0.07078 0.07314 0.9677 -0.07258 0.2141 fixed
sd_(Intercept).mother_pidlink 0.447 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8554 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.134 0.1942 -16.14 -3.514 -2.753 fixed
poly(age, 3, raw = TRUE)1 0.5388 0.03953 13.63 0.4614 0.6163 fixed
poly(age, 3, raw = TRUE)2 -0.0265 0.002502 -10.59 -0.0314 -0.02159 fixed
poly(age, 3, raw = TRUE)3 0.0004153 0.00004926 8.43 0.0003187 0.0005118 fixed
male -0.1777 0.02062 -8.619 -0.2182 -0.1373 fixed
count_birth_order2/2 0.05606 0.03495 1.604 -0.01245 0.1246 fixed
count_birth_order1/3 0.03115 0.04032 0.7725 -0.04788 0.1102 fixed
count_birth_order2/3 0.02824 0.0388 0.7279 -0.0478 0.1043 fixed
count_birth_order3/3 -0.01949 0.04253 -0.4582 -0.1028 0.06387 fixed
count_birth_order1/4 -0.082 0.06396 -1.282 -0.2074 0.04337 fixed
count_birth_order2/4 -0.06803 0.05601 -1.215 -0.1778 0.04174 fixed
count_birth_order3/4 -0.009767 0.05347 -0.1827 -0.1146 0.09503 fixed
count_birth_order4/4 -0.007113 0.05289 -0.1345 -0.1108 0.09656 fixed
count_birth_order1/5 -0.079 0.1123 -0.7037 -0.299 0.141 fixed
count_birth_order2/5 -0.0711 0.09979 -0.7125 -0.2667 0.1245 fixed
count_birth_order3/5 -0.07822 0.08421 -0.9289 -0.2433 0.08682 fixed
count_birth_order4/5 -0.1583 0.0694 -2.281 -0.2943 -0.0223 fixed
count_birth_order5/5 -0.192 0.06727 -2.855 -0.3239 -0.06018 fixed
count_birth_order1/5+ -0.1892 0.1439 -1.315 -0.4711 0.0928 fixed
count_birth_order2/5+ -0.1647 0.1316 -1.252 -0.4226 0.09312 fixed
count_birth_order3/5+ -0.2041 0.1084 -1.883 -0.4165 0.008378 fixed
count_birth_order4/5+ -0.3053 0.09716 -3.142 -0.4957 -0.1148 fixed
count_birth_order5/5+ -0.2496 0.07779 -3.208 -0.4021 -0.09711 fixed
count_birth_order5+/5+ -0.1526 0.05088 -2.999 -0.2523 -0.05286 fixed
sd_(Intercept).mother_pidlink 0.446 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8561 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 23286 23364 -11632 23264 NA NA NA
12 23288 23373 -11632 23264 0.04718 1 0.828
16 23292 23405 -11630 23260 4.367 4 0.3586
26 23305 23489 -11627 23253 6.595 10 0.7631

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.094 0.1905 -16.24 -3.467 -2.72 fixed
poly(age, 3, raw = TRUE)1 0.5341 0.03912 13.65 0.4575 0.6108 fixed
poly(age, 3, raw = TRUE)2 -0.02625 0.00248 -10.58 -0.03111 -0.02139 fixed
poly(age, 3, raw = TRUE)3 0.0004103 0.00004886 8.398 0.0003145 0.0005061 fixed
male -0.1813 0.02048 -8.849 -0.2214 -0.1411 fixed
sibling_count3 0.01962 0.02964 0.6619 -0.03847 0.0777 fixed
sibling_count4 -0.04539 0.03474 -1.307 -0.1135 0.02269 fixed
sibling_count5 -0.05441 0.0406 -1.34 -0.134 0.02517 fixed
sibling_count5+ -0.1553 0.03618 -4.292 -0.2262 -0.08437 fixed
sd_(Intercept).mother_pidlink 0.4484 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8547 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.063 0.1921 -15.94 -3.439 -2.686 fixed
birth_order -0.009473 0.007631 -1.241 -0.02443 0.005484 fixed
poly(age, 3, raw = TRUE)1 0.5313 0.03919 13.56 0.4544 0.6081 fixed
poly(age, 3, raw = TRUE)2 -0.02611 0.002483 -10.52 -0.03098 -0.02125 fixed
poly(age, 3, raw = TRUE)3 0.0004079 0.0000489 8.342 0.0003121 0.0005037 fixed
male -0.181 0.02048 -8.835 -0.2211 -0.1408 fixed
sibling_count3 0.02553 0.03001 0.8508 -0.03329 0.08435 fixed
sibling_count4 -0.0321 0.03634 -0.8832 -0.1033 0.03913 fixed
sibling_count5 -0.03351 0.04394 -0.7625 -0.1196 0.05262 fixed
sibling_count5+ -0.1126 0.04991 -2.256 -0.2105 -0.0148 fixed
sd_(Intercept).mother_pidlink 0.4477 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.855 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.096 0.192 -16.12 -3.473 -2.72 fixed
poly(age, 3, raw = TRUE)1 0.5334 0.03921 13.61 0.4566 0.6103 fixed
poly(age, 3, raw = TRUE)2 -0.02621 0.002483 -10.56 -0.03107 -0.02134 fixed
poly(age, 3, raw = TRUE)3 0.0004093 0.00004889 8.372 0.0003135 0.0005052 fixed
male -0.1816 0.02048 -8.865 -0.2217 -0.1414 fixed
sibling_count3 0.008918 0.03129 0.2851 -0.0524 0.07024 fixed
sibling_count4 -0.04393 0.03939 -1.115 -0.1211 0.03326 fixed
sibling_count5 -0.02658 0.04819 -0.5516 -0.121 0.06787 fixed
sibling_count5+ -0.1216 0.05324 -2.285 -0.226 -0.01729 fixed
birth_order_nonlinear2 0.01922 0.02567 0.7485 -0.0311 0.06954 fixed
birth_order_nonlinear3 0.03887 0.03402 1.143 -0.0278 0.1055 fixed
birth_order_nonlinear4 -0.03314 0.04453 -0.7443 -0.1204 0.05414 fixed
birth_order_nonlinear5 -0.07932 0.05598 -1.417 -0.189 0.03039 fixed
birth_order_nonlinear5+ -0.02667 0.06012 -0.4437 -0.1445 0.09115 fixed
sd_(Intercept).mother_pidlink 0.449 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8544 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.101 0.193 -16.06 -3.479 -2.722 fixed
poly(age, 3, raw = TRUE)1 0.5322 0.03927 13.55 0.4552 0.6092 fixed
poly(age, 3, raw = TRUE)2 -0.02612 0.002487 -10.51 -0.031 -0.02125 fixed
poly(age, 3, raw = TRUE)3 0.0004076 0.00004897 8.324 0.0003116 0.0005036 fixed
male -0.1814 0.02049 -8.852 -0.2216 -0.1412 fixed
count_birth_order2/2 0.04505 0.03813 1.182 -0.02967 0.1198 fixed
count_birth_order1/3 0.05846 0.0417 1.402 -0.02328 0.1402 fixed
count_birth_order2/3 0.03561 0.0408 0.8727 -0.04436 0.1156 fixed
count_birth_order3/3 0.009352 0.04477 0.2089 -0.07839 0.09709 fixed
count_birth_order1/4 -0.1275 0.0624 -2.043 -0.2498 -0.005167 fixed
count_birth_order2/4 -0.05352 0.05596 -0.9563 -0.1632 0.05616 fixed
count_birth_order3/4 0.04619 0.05355 0.8626 -0.05877 0.1512 fixed
count_birth_order4/4 -0.01379 0.05409 -0.2549 -0.1198 0.09222 fixed
count_birth_order1/5 0.04611 0.09122 0.5055 -0.1327 0.2249 fixed
count_birth_order2/5 -0.02527 0.08146 -0.3102 -0.1849 0.1344 fixed
count_birth_order3/5 0.07466 0.07305 1.022 -0.06851 0.2178 fixed
count_birth_order4/5 -0.09296 0.06719 -1.384 -0.2246 0.03873 fixed
count_birth_order5/5 -0.108 0.06459 -1.672 -0.2346 0.0186 fixed
count_birth_order1/5+ -0.1115 0.09781 -1.14 -0.3032 0.08019 fixed
count_birth_order2/5+ -0.04052 0.109 -0.3718 -0.2542 0.1731 fixed
count_birth_order3/5+ -0.05131 0.08401 -0.6108 -0.216 0.1133 fixed
count_birth_order4/5+ -0.2026 0.07702 -2.63 -0.3535 -0.05163 fixed
count_birth_order5/5+ -0.1804 0.06692 -2.697 -0.3116 -0.04929 fixed
count_birth_order5+/5+ -0.139 0.04611 -3.015 -0.2294 -0.04864 fixed
sd_(Intercept).mother_pidlink 0.4485 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8545 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 23568 23646 -11773 23546 NA NA NA
12 23568 23653 -11772 23544 1.543 1 0.2141
16 23572 23685 -11770 23540 4.59 4 0.332
26 23580 23763 -11764 23528 12.22 10 0.2709

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.076 0.1931 -15.93 -3.455 -2.698 fixed
poly(age, 3, raw = TRUE)1 0.5318 0.03963 13.42 0.4541 0.6095 fixed
poly(age, 3, raw = TRUE)2 -0.02602 0.00251 -10.37 -0.03094 -0.0211 fixed
poly(age, 3, raw = TRUE)3 0.0004051 0.0000494 8.201 0.0003083 0.0005019 fixed
male -0.1757 0.02083 -8.438 -0.2166 -0.1349 fixed
sibling_count3 -0.01308 0.02837 -0.461 -0.06868 0.04252 fixed
sibling_count4 -0.05899 0.03492 -1.689 -0.1274 0.00945 fixed
sibling_count5 -0.1589 0.04606 -3.449 -0.2491 -0.06859 fixed
sibling_count5+ -0.2128 0.04257 -4.998 -0.2962 -0.1293 fixed
sd_(Intercept).mother_pidlink 0.4509 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8514 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.092 0.1955 -15.81 -3.475 -2.708 fixed
birth_order 0.00463 0.009373 0.494 -0.01374 0.023 fixed
poly(age, 3, raw = TRUE)1 0.5333 0.03974 13.42 0.4554 0.6112 fixed
poly(age, 3, raw = TRUE)2 -0.02609 0.002514 -10.38 -0.03102 -0.02117 fixed
poly(age, 3, raw = TRUE)3 0.0004064 0.00004947 8.216 0.0003095 0.0005034 fixed
male -0.1758 0.02083 -8.439 -0.2166 -0.1349 fixed
sibling_count3 -0.01609 0.02902 -0.5545 -0.07296 0.04078 fixed
sibling_count4 -0.0657 0.03747 -1.753 -0.1392 0.007743 fixed
sibling_count5 -0.1699 0.05117 -3.32 -0.2702 -0.06958 fixed
sibling_count5+ -0.2345 0.06122 -3.83 -0.3545 -0.1145 fixed
sd_(Intercept).mother_pidlink 0.451 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8514 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.104 0.195 -15.92 -3.487 -2.722 fixed
poly(age, 3, raw = TRUE)1 0.535 0.03977 13.45 0.4571 0.613 fixed
poly(age, 3, raw = TRUE)2 -0.0262 0.002515 -10.42 -0.03113 -0.02127 fixed
poly(age, 3, raw = TRUE)3 0.0004084 0.00004949 8.251 0.0003114 0.0005054 fixed
male -0.1762 0.02083 -8.455 -0.217 -0.1353 fixed
sibling_count3 -0.01488 0.03034 -0.4904 -0.07435 0.04459 fixed
sibling_count4 -0.06779 0.04052 -1.673 -0.1472 0.01162 fixed
sibling_count5 -0.1548 0.05619 -2.754 -0.2649 -0.04464 fixed
sibling_count5+ -0.2456 0.06644 -3.696 -0.3758 -0.1153 fixed
birth_order_nonlinear2 0.02627 0.02496 1.052 -0.02265 0.0752 fixed
birth_order_nonlinear3 0.003957 0.03499 0.1131 -0.06463 0.07254 fixed
birth_order_nonlinear4 0.03323 0.04852 0.6848 -0.06187 0.1283 fixed
birth_order_nonlinear5 -0.02669 0.06674 -0.3999 -0.1575 0.1041 fixed
birth_order_nonlinear5+ 0.07335 0.07557 0.9707 -0.07476 0.2215 fixed
sd_(Intercept).mother_pidlink 0.4505 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8517 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.108 0.1959 -15.87 -3.492 -2.724 fixed
poly(age, 3, raw = TRUE)1 0.5337 0.03987 13.39 0.4556 0.6119 fixed
poly(age, 3, raw = TRUE)2 -0.02612 0.002521 -10.36 -0.03107 -0.02118 fixed
poly(age, 3, raw = TRUE)3 0.0004071 0.00004962 8.205 0.0003098 0.0005043 fixed
male -0.1755 0.02085 -8.42 -0.2164 -0.1347 fixed
count_birth_order2/2 0.05598 0.0346 1.618 -0.01183 0.1238 fixed
count_birth_order1/3 0.0261 0.04051 0.6443 -0.0533 0.1055 fixed
count_birth_order2/3 0.01806 0.03892 0.464 -0.05822 0.09434 fixed
count_birth_order3/3 -0.02986 0.04272 -0.6989 -0.1136 0.05387 fixed
count_birth_order1/4 -0.09554 0.06434 -1.485 -0.2216 0.03056 fixed
count_birth_order2/4 -0.0866 0.05675 -1.526 -0.1978 0.02463 fixed
count_birth_order3/4 -0.0142 0.05387 -0.2637 -0.1198 0.09137 fixed
count_birth_order4/4 0.009754 0.05365 0.1818 -0.0954 0.1149 fixed
count_birth_order1/5 -0.05029 0.1121 -0.4487 -0.2699 0.1694 fixed
count_birth_order2/5 -0.1172 0.1071 -1.094 -0.3272 0.09276 fixed
count_birth_order3/5 -0.1053 0.08641 -1.218 -0.2747 0.06408 fixed
count_birth_order4/5 -0.123 0.07302 -1.685 -0.2661 0.02011 fixed
count_birth_order5/5 -0.2081 0.0713 -2.919 -0.3479 -0.0684 fixed
count_birth_order1/5+ -0.2608 0.145 -1.799 -0.545 0.02333 fixed
count_birth_order2/5+ -0.1588 0.1336 -1.188 -0.4207 0.1031 fixed
count_birth_order3/5+ -0.2241 0.1089 -2.058 -0.4375 -0.01066 fixed
count_birth_order4/5+ -0.3077 0.1033 -2.979 -0.5102 -0.1052 fixed
count_birth_order5/5+ -0.2128 0.08176 -2.602 -0.373 -0.05251 fixed
count_birth_order5+/5+ -0.1609 0.05288 -3.043 -0.2646 -0.05726 fixed
sd_(Intercept).mother_pidlink 0.4492 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8523 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 22637 22714 -11307 22615 NA NA NA
12 22638 22723 -11307 22614 0.2439 1 0.6214
16 22643 22756 -11306 22611 3.053 4 0.549
26 22655 22837 -11302 22603 8.409 10 0.5889

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Raven 2007 old

birthorder <- birthorder %>% mutate(outcome = raven_2007_old)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
             data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -5.555 1.991 -2.79 -9.456 -1.653 fixed
poly(age, 3, raw = TRUE)1 0.62 0.207 2.996 0.2143 1.026 fixed
poly(age, 3, raw = TRUE)2 -0.02159 0.007077 -3.05 -0.03546 -0.007718 fixed
poly(age, 3, raw = TRUE)3 0.0002391 0.00007963 3.003 0.00008304 0.0003952 fixed
male 0.1589 0.02117 7.507 0.1174 0.2004 fixed
sibling_count3 -0.02469 0.05681 -0.4346 -0.136 0.08666 fixed
sibling_count4 -0.1125 0.05551 -2.027 -0.2213 -0.003706 fixed
sibling_count5 -0.02214 0.05634 -0.393 -0.1326 0.08828 fixed
sibling_count5+ -0.252 0.0464 -5.432 -0.343 -0.1611 fixed
sd_(Intercept).mother_pidlink 0.5145 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8407 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -5.54 1.991 -2.782 -9.443 -1.637 fixed
birth_order -0.001469 0.004356 -0.3373 -0.01001 0.007068 fixed
poly(age, 3, raw = TRUE)1 0.6189 0.207 2.99 0.2132 1.025 fixed
poly(age, 3, raw = TRUE)2 -0.02156 0.007078 -3.045 -0.03543 -0.007682 fixed
poly(age, 3, raw = TRUE)3 0.0002387 0.00007964 2.997 0.00008259 0.0003948 fixed
male 0.159 0.02117 7.511 0.1175 0.2005 fixed
sibling_count3 -0.02407 0.05684 -0.4234 -0.1355 0.08734 fixed
sibling_count4 -0.1112 0.05564 -1.998 -0.2203 -0.002142 fixed
sibling_count5 -0.01996 0.05671 -0.3519 -0.1311 0.0912 fixed
sibling_count5+ -0.2462 0.0495 -4.974 -0.3433 -0.1492 fixed
sd_(Intercept).mother_pidlink 0.5144 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8408 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -5.504 1.992 -2.763 -9.409 -1.599 fixed
poly(age, 3, raw = TRUE)1 0.6189 0.2071 2.989 0.213 1.025 fixed
poly(age, 3, raw = TRUE)2 -0.02158 0.007082 -3.047 -0.03546 -0.007696 fixed
poly(age, 3, raw = TRUE)3 0.0002393 0.00007968 3.003 0.00008312 0.0003955 fixed
male 0.158 0.02117 7.464 0.1165 0.1995 fixed
sibling_count3 -0.0224 0.05765 -0.3885 -0.1354 0.0906 fixed
sibling_count4 -0.1007 0.05684 -1.772 -0.2121 0.01067 fixed
sibling_count5 -0.007411 0.0584 -0.1269 -0.1219 0.107 fixed
sibling_count5+ -0.2452 0.05113 -4.796 -0.3454 -0.145 fixed
birth_order_nonlinear2 -0.08067 0.03356 -2.404 -0.1464 -0.01489 fixed
birth_order_nonlinear3 -0.04792 0.03615 -1.325 -0.1188 0.02294 fixed
birth_order_nonlinear4 -0.07509 0.03956 -1.898 -0.1526 0.002444 fixed
birth_order_nonlinear5 -0.05645 0.04473 -1.262 -0.1441 0.03123 fixed
birth_order_nonlinear5+ -0.03028 0.0381 -0.7947 -0.1049 0.04439 fixed
sd_(Intercept).mother_pidlink 0.515 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8403 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -5.585 1.994 -2.801 -9.492 -1.677 fixed
poly(age, 3, raw = TRUE)1 0.629 0.2072 3.035 0.2228 1.035 fixed
poly(age, 3, raw = TRUE)2 -0.02193 0.007086 -3.095 -0.03582 -0.008044 fixed
poly(age, 3, raw = TRUE)3 0.0002436 0.00007972 3.055 0.0000873 0.0003998 fixed
male 0.1577 0.02118 7.445 0.1162 0.1992 fixed
count_birth_order2/2 -0.1162 0.07779 -1.493 -0.2686 0.0363 fixed
count_birth_order1/3 -0.02636 0.07751 -0.34 -0.1783 0.1256 fixed
count_birth_order2/3 -0.07158 0.0828 -0.8644 -0.2339 0.09071 fixed
count_birth_order3/3 -0.1474 0.08305 -1.775 -0.3102 0.01532 fixed
count_birth_order1/4 -0.05173 0.07984 -0.6479 -0.2082 0.1048 fixed
count_birth_order2/4 -0.1402 0.08328 -1.684 -0.3035 0.02298 fixed
count_birth_order3/4 -0.3107 0.08564 -3.628 -0.4786 -0.1429 fixed
count_birth_order4/4 -0.2031 0.08645 -2.349 -0.3725 -0.03362 fixed
count_birth_order1/5 -0.05713 0.08748 -0.653 -0.2286 0.1143 fixed
count_birth_order2/5 -0.18 0.08917 -2.018 -0.3547 -0.005209 fixed
count_birth_order3/5 -0.0007824 0.08803 -0.008888 -0.1733 0.1717 fixed
count_birth_order4/5 -0.1157 0.09013 -1.284 -0.2924 0.06092 fixed
count_birth_order5/5 -0.01407 0.0919 -0.1531 -0.1942 0.166 fixed
count_birth_order1/5+ -0.2993 0.06996 -4.279 -0.4365 -0.1622 fixed
count_birth_order2/5+ -0.3558 0.07041 -5.053 -0.4938 -0.2178 fixed
count_birth_order3/5+ -0.2488 0.06923 -3.594 -0.3845 -0.1131 fixed
count_birth_order4/5+ -0.3295 0.06798 -4.847 -0.4627 -0.1962 fixed
count_birth_order5/5+ -0.3357 0.06877 -4.881 -0.4705 -0.2009 fixed
count_birth_order5+/5+ -0.2904 0.06009 -4.832 -0.4082 -0.1726 fixed
sd_(Intercept).mother_pidlink 0.5126 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8409 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 22194 22271 -11086 22172 NA NA NA
12 22196 22280 -11086 22172 0.1146 1 0.735
16 22197 22309 -11083 22165 7.31 4 0.1204
26 22200 22382 -11074 22148 16.95 10 0.07548

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -5.623 3.55 -1.584 -12.58 1.335 fixed
poly(age, 3, raw = TRUE)1 0.6273 0.3694 1.698 -0.09661 1.351 fixed
poly(age, 3, raw = TRUE)2 -0.02163 0.01263 -1.713 -0.04638 0.003113 fixed
poly(age, 3, raw = TRUE)3 0.0002424 0.0001419 1.708 -0.00003571 0.0005205 fixed
male 0.1013 0.02818 3.595 0.04608 0.1565 fixed
sibling_count3 -0.0009628 0.05098 -0.01889 -0.1009 0.09895 fixed
sibling_count4 -0.1017 0.05264 -1.933 -0.2049 0.001433 fixed
sibling_count5 -0.1673 0.05745 -2.911 -0.2798 -0.05466 fixed
sibling_count5+ -0.285 0.04968 -5.736 -0.3824 -0.1876 fixed
sd_(Intercept).mother_pidlink 0.4319 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7512 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -5.809 3.549 -1.637 -12.76 1.147 fixed
birth_order 0.01696 0.009109 1.862 -0.0008961 0.03481 fixed
poly(age, 3, raw = TRUE)1 0.6437 0.3692 1.743 -0.07996 1.367 fixed
poly(age, 3, raw = TRUE)2 -0.02221 0.01262 -1.76 -0.04695 0.002527 fixed
poly(age, 3, raw = TRUE)3 0.0002496 0.0001418 1.759 -0.00002846 0.0005276 fixed
male 0.09962 0.02818 3.535 0.04439 0.1548 fixed
sibling_count3 -0.008611 0.05115 -0.1683 -0.1089 0.09164 fixed
sibling_count4 -0.1211 0.05368 -2.256 -0.2263 -0.01587 fixed
sibling_count5 -0.1989 0.05992 -3.319 -0.3163 -0.08145 fixed
sibling_count5+ -0.3465 0.05965 -5.809 -0.4634 -0.2296 fixed
sd_(Intercept).mother_pidlink 0.4335 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7501 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.094 3.552 -1.716 -13.06 0.8668 fixed
poly(age, 3, raw = TRUE)1 0.6757 0.3695 1.829 -0.04856 1.4 fixed
poly(age, 3, raw = TRUE)2 -0.02331 0.01263 -1.845 -0.04807 0.001449 fixed
poly(age, 3, raw = TRUE)3 0.000262 0.000142 1.846 -0.0000162 0.0005402 fixed
male 0.09972 0.02816 3.542 0.04453 0.1549 fixed
sibling_count3 -0.0143 0.05181 -0.2761 -0.1159 0.08725 fixed
sibling_count4 -0.1186 0.05549 -2.137 -0.2273 -0.009797 fixed
sibling_count5 -0.1841 0.06264 -2.939 -0.3069 -0.06134 fixed
sibling_count5+ -0.3681 0.06103 -6.031 -0.4877 -0.2484 fixed
birth_order_nonlinear2 0.001955 0.03711 0.05268 -0.07078 0.07469 fixed
birth_order_nonlinear3 0.05547 0.04397 1.261 -0.03072 0.1417 fixed
birth_order_nonlinear4 0.001765 0.05378 0.03282 -0.1036 0.1072 fixed
birth_order_nonlinear5 0.006667 0.06523 0.1022 -0.1212 0.1345 fixed
birth_order_nonlinear5+ 0.1875 0.06536 2.868 0.05937 0.3156 fixed
sd_(Intercept).mother_pidlink 0.434 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7494 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.348 3.556 -1.785 -13.32 0.6221 fixed
poly(age, 3, raw = TRUE)1 0.7056 0.37 1.907 -0.01961 1.431 fixed
poly(age, 3, raw = TRUE)2 -0.02436 0.01265 -1.926 -0.04916 0.0004275 fixed
poly(age, 3, raw = TRUE)3 0.0002743 0.0001421 1.929 -0.000004341 0.0005529 fixed
male 0.1005 0.02819 3.564 0.04522 0.1557 fixed
count_birth_order2/2 -0.06828 0.07074 -0.9652 -0.2069 0.07037 fixed
count_birth_order1/3 -0.05194 0.06509 -0.798 -0.1795 0.07564 fixed
count_birth_order2/3 0.02114 0.0713 0.2966 -0.1186 0.1609 fixed
count_birth_order3/3 -0.04072 0.07914 -0.5146 -0.1958 0.1144 fixed
count_birth_order1/4 -0.1329 0.07592 -1.751 -0.2817 0.01585 fixed
count_birth_order2/4 -0.1618 0.0776 -2.086 -0.3139 -0.009757 fixed
count_birth_order3/4 -0.1453 0.07867 -1.847 -0.2995 0.00886 fixed
count_birth_order4/4 -0.06372 0.0855 -0.7453 -0.2313 0.1039 fixed
count_birth_order1/5 -0.2621 0.09772 -2.682 -0.4536 -0.07053 fixed
count_birth_order2/5 -0.282 0.1013 -2.785 -0.4805 -0.08352 fixed
count_birth_order3/5 -0.02152 0.09263 -0.2323 -0.2031 0.16 fixed
count_birth_order4/5 -0.2723 0.09185 -2.965 -0.4523 -0.09228 fixed
count_birth_order5/5 -0.154 0.09646 -1.596 -0.343 0.03509 fixed
count_birth_order1/5+ -0.4131 0.09369 -4.409 -0.5967 -0.2294 fixed
count_birth_order2/5+ -0.3539 0.09094 -3.892 -0.5322 -0.1757 fixed
count_birth_order3/5+ -0.2987 0.08559 -3.489 -0.4664 -0.1309 fixed
count_birth_order4/5+ -0.4136 0.08515 -4.857 -0.5805 -0.2467 fixed
count_birth_order5/5+ -0.4167 0.08265 -5.042 -0.5787 -0.2547 fixed
count_birth_order5+/5+ -0.2074 0.06516 -3.183 -0.3351 -0.07967 fixed
sd_(Intercept).mother_pidlink 0.4347 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7488 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 9087 9155 -4533 9065 NA NA NA
12 9086 9160 -4531 9062 3.462 1 0.06281
16 9086 9185 -4527 9054 8.221 4 0.08383
26 9094 9255 -4521 9042 11.85 10 0.2952

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -5.806 3.542 -1.639 -12.75 1.135 fixed
poly(age, 3, raw = TRUE)1 0.649 0.3685 1.761 -0.07334 1.371 fixed
poly(age, 3, raw = TRUE)2 -0.02239 0.0126 -1.778 -0.04709 0.002298 fixed
poly(age, 3, raw = TRUE)3 0.0002508 0.0001416 1.772 -0.00002667 0.0005283 fixed
male 0.1003 0.0281 3.569 0.04521 0.1554 fixed
sibling_count3 -0.04235 0.05637 -0.7513 -0.1528 0.06813 fixed
sibling_count4 -0.07353 0.05708 -1.288 -0.1854 0.03834 fixed
sibling_count5 -0.1384 0.05942 -2.329 -0.2549 -0.02194 fixed
sibling_count5+ -0.2571 0.05173 -4.97 -0.3585 -0.1557 fixed
sd_(Intercept).mother_pidlink 0.4344 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7502 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -5.883 3.543 -1.661 -12.83 1.061 fixed
birth_order 0.006245 0.008017 0.779 -0.009468 0.02196 fixed
poly(age, 3, raw = TRUE)1 0.656 0.3686 1.78 -0.06646 1.379 fixed
poly(age, 3, raw = TRUE)2 -0.02265 0.0126 -1.797 -0.04734 0.002052 fixed
poly(age, 3, raw = TRUE)3 0.0002539 0.0001416 1.793 -0.00002362 0.0005315 fixed
male 0.09972 0.02811 3.548 0.04463 0.1548 fixed
sibling_count3 -0.04529 0.05651 -0.8015 -0.1561 0.06547 fixed
sibling_count4 -0.08009 0.05772 -1.388 -0.1932 0.03304 fixed
sibling_count5 -0.1488 0.06093 -2.443 -0.2683 -0.02942 fixed
sibling_count5+ -0.2793 0.05906 -4.729 -0.3951 -0.1635 fixed
sd_(Intercept).mother_pidlink 0.4351 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7499 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.038 3.547 -1.702 -12.99 0.9143 fixed
poly(age, 3, raw = TRUE)1 0.6738 0.3691 1.826 -0.04963 1.397 fixed
poly(age, 3, raw = TRUE)2 -0.02326 0.01262 -1.843 -0.04799 0.001471 fixed
poly(age, 3, raw = TRUE)3 0.0002609 0.0001418 1.84 -0.000017 0.0005388 fixed
male 0.09838 0.02811 3.499 0.04327 0.1535 fixed
sibling_count3 -0.04419 0.05725 -0.7718 -0.1564 0.06802 fixed
sibling_count4 -0.07128 0.05925 -1.203 -0.1874 0.04485 fixed
sibling_count5 -0.1331 0.06322 -2.105 -0.257 -0.009199 fixed
sibling_count5+ -0.2882 0.06054 -4.76 -0.4068 -0.1695 fixed
birth_order_nonlinear2 -0.01585 0.03793 -0.4179 -0.09019 0.05849 fixed
birth_order_nonlinear3 0.002151 0.04391 0.04899 -0.08391 0.08822 fixed
birth_order_nonlinear4 -0.02857 0.05246 -0.5446 -0.1314 0.07425 fixed
birth_order_nonlinear5 -0.02348 0.0626 -0.3751 -0.1462 0.09921 fixed
birth_order_nonlinear5+ 0.08848 0.05961 1.484 -0.02837 0.2053 fixed
sd_(Intercept).mother_pidlink 0.437 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.749 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.25 3.558 -1.757 -13.22 0.7235 fixed
poly(age, 3, raw = TRUE)1 0.6961 0.3702 1.88 -0.02951 1.422 fixed
poly(age, 3, raw = TRUE)2 -0.02407 0.01265 -1.902 -0.04887 0.0007274 fixed
poly(age, 3, raw = TRUE)3 0.0002707 0.0001422 1.904 -0.000007981 0.0005494 fixed
male 0.09831 0.02817 3.49 0.0431 0.1535 fixed
count_birth_order2/2 0.01898 0.08003 0.2371 -0.1379 0.1758 fixed
count_birth_order1/3 -0.0516 0.07245 -0.7123 -0.1936 0.09039 fixed
count_birth_order2/3 0.01223 0.07875 0.1553 -0.1421 0.1666 fixed
count_birth_order3/3 -0.06923 0.08565 -0.8082 -0.2371 0.09865 fixed
count_birth_order1/4 -0.00921 0.08039 -0.1146 -0.1668 0.1483 fixed
count_birth_order2/4 -0.1153 0.08129 -1.418 -0.2746 0.04404 fixed
count_birth_order3/4 -0.1334 0.08583 -1.554 -0.3016 0.03488 fixed
count_birth_order4/4 -0.01166 0.09462 -0.1233 -0.1971 0.1738 fixed
count_birth_order1/5 -0.1156 0.09221 -1.254 -0.2964 0.06509 fixed
count_birth_order2/5 -0.1886 0.09582 -1.968 -0.3764 -0.0008033 fixed
count_birth_order3/5 -0.03245 0.09584 -0.3386 -0.2203 0.1554 fixed
count_birth_order4/5 -0.2382 0.09614 -2.478 -0.4267 -0.04981 fixed
count_birth_order5/5 -0.09026 0.09993 -0.9032 -0.2861 0.1056 fixed
count_birth_order1/5+ -0.2836 0.08558 -3.314 -0.4514 -0.1159 fixed
count_birth_order2/5+ -0.3124 0.08718 -3.583 -0.4833 -0.1415 fixed
count_birth_order3/5+ -0.2233 0.08297 -2.692 -0.3859 -0.06071 fixed
count_birth_order4/5+ -0.3001 0.08217 -3.653 -0.4612 -0.1391 fixed
count_birth_order5/5+ -0.3303 0.08399 -3.933 -0.4949 -0.1657 fixed
count_birth_order5+/5+ -0.1878 0.06719 -2.795 -0.3195 -0.0561 fixed
sd_(Intercept).mother_pidlink 0.4372 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7489 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 9142 9210 -4560 9120 NA NA NA
12 9143 9217 -4560 9119 0.6053 1 0.4366
16 9147 9246 -4558 9115 4.017 4 0.4037
26 9157 9318 -4552 9105 10.13 10 0.4292

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.585 3.572 -1.843 -13.59 0.4161 fixed
poly(age, 3, raw = TRUE)1 0.7268 0.3715 1.956 -0.001401 1.455 fixed
poly(age, 3, raw = TRUE)2 -0.02503 0.01269 -1.971 -0.04991 -0.0001444 fixed
poly(age, 3, raw = TRUE)3 0.00028 0.0001426 1.963 0.0000004695 0.0005594 fixed
male 0.0975 0.02838 3.435 0.04187 0.1531 fixed
sibling_count3 0.0009891 0.05023 0.01969 -0.09747 0.09945 fixed
sibling_count4 -0.07573 0.0521 -1.454 -0.1778 0.02638 fixed
sibling_count5 -0.1448 0.05821 -2.489 -0.2589 -0.03077 fixed
sibling_count5+ -0.2645 0.04975 -5.316 -0.362 -0.167 fixed
sd_(Intercept).mother_pidlink 0.4376 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7484 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.765 3.572 -1.894 -13.77 0.2352 fixed
birth_order 0.01577 0.009347 1.687 -0.002547 0.03409 fixed
poly(age, 3, raw = TRUE)1 0.7427 0.3715 1.999 0.01464 1.471 fixed
poly(age, 3, raw = TRUE)2 -0.02558 0.01269 -2.016 -0.05046 -0.0007059 fixed
poly(age, 3, raw = TRUE)3 0.0002869 0.0001426 2.012 0.000007389 0.0005663 fixed
male 0.09624 0.02838 3.391 0.04062 0.1519 fixed
sibling_count3 -0.006248 0.05042 -0.1239 -0.1051 0.09258 fixed
sibling_count4 -0.09368 0.05318 -1.761 -0.1979 0.01056 fixed
sibling_count5 -0.173 0.06055 -2.856 -0.2916 -0.05428 fixed
sibling_count5+ -0.3216 0.06016 -5.346 -0.4395 -0.2037 fixed
sd_(Intercept).mother_pidlink 0.4387 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7476 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -7.133 3.572 -1.997 -14.13 -0.1313 fixed
poly(age, 3, raw = TRUE)1 0.7832 0.3715 2.108 0.055 1.511 fixed
poly(age, 3, raw = TRUE)2 -0.02698 0.0127 -2.126 -0.05187 -0.002102 fixed
poly(age, 3, raw = TRUE)3 0.0003029 0.0001426 2.124 0.00002336 0.0005824 fixed
male 0.09676 0.02835 3.413 0.0412 0.1523 fixed
sibling_count3 -0.01456 0.05114 -0.2848 -0.1148 0.08568 fixed
sibling_count4 -0.09135 0.05505 -1.659 -0.1992 0.01655 fixed
sibling_count5 -0.1603 0.06309 -2.541 -0.284 -0.03666 fixed
sibling_count5+ -0.3484 0.06163 -5.652 -0.4692 -0.2276 fixed
birth_order_nonlinear2 0.005591 0.03694 0.1513 -0.06682 0.078 fixed
birth_order_nonlinear3 0.06332 0.04377 1.446 -0.02248 0.1491 fixed
birth_order_nonlinear4 -0.009146 0.05455 -0.1677 -0.1161 0.09776 fixed
birth_order_nonlinear5 0.00588 0.06736 0.08729 -0.1261 0.1379 fixed
birth_order_nonlinear5+ 0.1927 0.06732 2.862 0.06073 0.3246 fixed
sd_(Intercept).mother_pidlink 0.4399 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7463 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -7.193 3.574 -2.013 -14.2 -0.1878 fixed
poly(age, 3, raw = TRUE)1 0.7935 0.3717 2.134 0.06486 1.522 fixed
poly(age, 3, raw = TRUE)2 -0.02739 0.0127 -2.156 -0.05229 -0.002496 fixed
poly(age, 3, raw = TRUE)3 0.0003081 0.0001427 2.159 0.0000284 0.0005877 fixed
male 0.0977 0.02838 3.442 0.04207 0.1533 fixed
count_birth_order2/2 -0.05581 0.06848 -0.815 -0.19 0.0784 fixed
count_birth_order1/3 -0.04644 0.06417 -0.7237 -0.1722 0.07932 fixed
count_birth_order2/3 0.01644 0.071 0.2316 -0.1227 0.1556 fixed
count_birth_order3/3 -0.01701 0.07702 -0.2208 -0.168 0.1339 fixed
count_birth_order1/4 -0.09631 0.07556 -1.275 -0.2444 0.05178 fixed
count_birth_order2/4 -0.1318 0.0772 -1.708 -0.2831 0.01948 fixed
count_birth_order3/4 -0.09476 0.0785 -1.207 -0.2486 0.05911 fixed
count_birth_order4/4 -0.068 0.08417 -0.8078 -0.233 0.09698 fixed
count_birth_order1/5 -0.2536 0.0969 -2.617 -0.4435 -0.06366 fixed
count_birth_order2/5 -0.3178 0.1027 -3.096 -0.5191 -0.1166 fixed
count_birth_order3/5 0.04751 0.09371 0.507 -0.1362 0.2312 fixed
count_birth_order4/5 -0.2302 0.09418 -2.444 -0.4148 -0.04563 fixed
count_birth_order5/5 -0.1075 0.1022 -1.051 -0.3078 0.09286 fixed
count_birth_order1/5+ -0.3875 0.09583 -4.044 -0.5753 -0.1997 fixed
count_birth_order2/5+ -0.2473 0.09394 -2.633 -0.4314 -0.06318 fixed
count_birth_order3/5+ -0.3206 0.08569 -3.741 -0.4885 -0.1526 fixed
count_birth_order4/5+ -0.4067 0.08779 -4.633 -0.5788 -0.2347 fixed
count_birth_order5/5+ -0.4028 0.08341 -4.83 -0.5663 -0.2394 fixed
count_birth_order5+/5+ -0.1803 0.06577 -2.742 -0.3092 -0.05144 fixed
sd_(Intercept).mother_pidlink 0.4408 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7452 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 8931 8999 -4454 8909 NA NA NA
12 8930 9004 -4453 8906 2.848 1 0.09148
16 8928 9027 -4448 8896 9.558 4 0.04856
26 8933 9093 -4440 8881 15.3 10 0.1215

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Numeracy 2007 old

birthorder <- birthorder %>% mutate(outcome = math_2007_old)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
             data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -5.1 1.962 -2.599 -8.946 -1.254 fixed
poly(age, 3, raw = TRUE)1 0.6087 0.2039 2.985 0.209 1.008 fixed
poly(age, 3, raw = TRUE)2 -0.02176 0.006973 -3.12 -0.03543 -0.00809 fixed
poly(age, 3, raw = TRUE)3 0.0002442 0.00007845 3.114 0.00009049 0.000398 fixed
male -0.1091 0.02082 -5.241 -0.1499 -0.06829 fixed
sibling_count3 0.03799 0.05713 0.6651 -0.07397 0.15 fixed
sibling_count4 -0.1154 0.0559 -2.065 -0.225 -0.00586 fixed
sibling_count5 -0.0517 0.05682 -0.9099 -0.1631 0.05967 fixed
sibling_count5+ -0.2327 0.0467 -4.983 -0.3243 -0.1412 fixed
sd_(Intercept).mother_pidlink 0.5578 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8105 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -5.077 1.963 -2.587 -8.924 -1.231 fixed
birth_order -0.002346 0.004368 -0.5372 -0.01091 0.006214 fixed
poly(age, 3, raw = TRUE)1 0.6072 0.204 2.977 0.2074 1.007 fixed
poly(age, 3, raw = TRUE)2 -0.02171 0.006974 -3.113 -0.03538 -0.008042 fixed
poly(age, 3, raw = TRUE)3 0.0002436 0.00007846 3.105 0.00008987 0.0003974 fixed
male -0.1089 0.02082 -5.233 -0.1497 -0.06813 fixed
sibling_count3 0.03898 0.05716 0.682 -0.07304 0.151 fixed
sibling_count4 -0.1133 0.05603 -2.023 -0.2231 -0.003524 fixed
sibling_count5 -0.0482 0.05719 -0.8428 -0.1603 0.06389 fixed
sibling_count5+ -0.2235 0.04977 -4.49 -0.321 -0.1259 fixed
sd_(Intercept).mother_pidlink 0.5576 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8106 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -5.042 1.964 -2.567 -8.892 -1.192 fixed
poly(age, 3, raw = TRUE)1 0.6034 0.2042 2.955 0.2032 1.004 fixed
poly(age, 3, raw = TRUE)2 -0.02158 0.006981 -3.091 -0.03526 -0.007896 fixed
poly(age, 3, raw = TRUE)3 0.0002423 0.00007853 3.086 0.0000884 0.0003962 fixed
male -0.1088 0.02083 -5.225 -0.1497 -0.06801 fixed
sibling_count3 0.04522 0.05794 0.7805 -0.06835 0.1588 fixed
sibling_count4 -0.1125 0.05717 -1.968 -0.2245 -0.0004324 fixed
sibling_count5 -0.0471 0.0588 -0.801 -0.1623 0.06814 fixed
sibling_count5+ -0.2317 0.05131 -4.515 -0.3322 -0.1311 fixed
birth_order_nonlinear2 -0.01638 0.03294 -0.4972 -0.08093 0.04818 fixed
birth_order_nonlinear3 -0.03209 0.03544 -0.9054 -0.1016 0.03738 fixed
birth_order_nonlinear4 0.003617 0.03882 0.09317 -0.07248 0.07971 fixed
birth_order_nonlinear5 -0.01445 0.04397 -0.3287 -0.1006 0.07172 fixed
birth_order_nonlinear5+ -0.004172 0.03781 -0.1103 -0.07828 0.06994 fixed
sd_(Intercept).mother_pidlink 0.5579 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8107 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -5.086 1.967 -2.586 -8.941 -1.231 fixed
poly(age, 3, raw = TRUE)1 0.6108 0.2044 2.988 0.2102 1.011 fixed
poly(age, 3, raw = TRUE)2 -0.02183 0.006989 -3.124 -0.03553 -0.008132 fixed
poly(age, 3, raw = TRUE)3 0.0002452 0.00007861 3.119 0.00009108 0.0003992 fixed
male -0.1091 0.02085 -5.235 -0.15 -0.06828 fixed
count_birth_order2/2 -0.08004 0.07646 -1.047 -0.2299 0.06981 fixed
count_birth_order1/3 0.00667 0.07716 0.08644 -0.1446 0.1579 fixed
count_birth_order2/3 0.04806 0.08229 0.5841 -0.1132 0.2094 fixed
count_birth_order3/3 -0.04683 0.08257 -0.5672 -0.2087 0.115 fixed
count_birth_order1/4 -0.1732 0.07947 -2.179 -0.3289 -0.01741 fixed
count_birth_order2/4 -0.133 0.08273 -1.607 -0.2951 0.02917 fixed
count_birth_order3/4 -0.2135 0.0851 -2.509 -0.3803 -0.04674 fixed
count_birth_order4/4 -0.08108 0.08588 -0.9441 -0.2494 0.08724 fixed
count_birth_order1/5 -0.06339 0.08697 -0.7288 -0.2338 0.1071 fixed
count_birth_order2/5 -0.1507 0.08848 -1.703 -0.3241 0.02274 fixed
count_birth_order3/5 -0.08961 0.08732 -1.026 -0.2607 0.08153 fixed
count_birth_order4/5 -0.1034 0.08935 -1.157 -0.2785 0.07172 fixed
count_birth_order5/5 -0.02592 0.09116 -0.2843 -0.2046 0.1527 fixed
count_birth_order1/5+ -0.258 0.0695 -3.712 -0.3942 -0.1218 fixed
count_birth_order2/5+ -0.2725 0.06994 -3.896 -0.4095 -0.1354 fixed
count_birth_order3/5+ -0.2665 0.06871 -3.879 -0.4012 -0.1318 fixed
count_birth_order4/5+ -0.2656 0.06748 -3.937 -0.3979 -0.1334 fixed
count_birth_order5/5+ -0.2939 0.06828 -4.304 -0.4277 -0.1601 fixed
count_birth_order5+/5+ -0.2644 0.05998 -4.408 -0.382 -0.1468 fixed
sd_(Intercept).mother_pidlink 0.5574 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8111 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 21998 22075 -10988 21976 NA NA NA
12 22000 22084 -10988 21976 0.2897 1 0.5904
16 22007 22119 -10988 21975 0.9632 4 0.9153
26 22021 22203 -10984 21969 6.232 10 0.7954

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -4.127 4.008 -1.03 -11.98 3.729 fixed
poly(age, 3, raw = TRUE)1 0.5153 0.417 1.236 -0.302 1.333 fixed
poly(age, 3, raw = TRUE)2 -0.01842 0.01425 -1.292 -0.04636 0.009517 fixed
poly(age, 3, raw = TRUE)3 0.0002076 0.0001602 1.296 -0.0001064 0.0005215 fixed
male -0.1656 0.0319 -5.192 -0.2281 -0.1031 fixed
sibling_count3 0.02809 0.0588 0.4777 -0.08715 0.1433 fixed
sibling_count4 -0.04592 0.06077 -0.7557 -0.165 0.07318 fixed
sibling_count5 -0.1236 0.06645 -1.86 -0.2539 0.006609 fixed
sibling_count5+ -0.268 0.05746 -4.664 -0.3806 -0.1554 fixed
sd_(Intercept).mother_pidlink 0.5375 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8326 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -4.343 4.008 -1.084 -12.2 3.512 fixed
birth_order 0.02037 0.01042 1.954 -0.00006569 0.0408 fixed
poly(age, 3, raw = TRUE)1 0.5341 0.417 1.281 -0.2832 1.351 fixed
poly(age, 3, raw = TRUE)2 -0.01908 0.01425 -1.338 -0.04701 0.008858 fixed
poly(age, 3, raw = TRUE)3 0.0002158 0.0001602 1.347 -0.00009812 0.0005297 fixed
male -0.1676 0.0319 -5.254 -0.2301 -0.1051 fixed
sibling_count3 0.01892 0.05896 0.3209 -0.09664 0.1345 fixed
sibling_count4 -0.06944 0.06192 -1.121 -0.1908 0.05193 fixed
sibling_count5 -0.162 0.06927 -2.339 -0.2978 -0.02625 fixed
sibling_count5+ -0.3421 0.06883 -4.97 -0.477 -0.2072 fixed
sd_(Intercept).mother_pidlink 0.5372 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8323 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -4.234 4.016 -1.054 -12.11 3.638 fixed
poly(age, 3, raw = TRUE)1 0.5252 0.4179 1.257 -0.2938 1.344 fixed
poly(age, 3, raw = TRUE)2 -0.01878 0.01428 -1.315 -0.04678 0.009209 fixed
poly(age, 3, raw = TRUE)3 0.0002127 0.0001605 1.325 -0.0001019 0.0005272 fixed
male -0.1671 0.03191 -5.237 -0.2297 -0.1046 fixed
sibling_count3 0.01475 0.05971 0.2471 -0.1023 0.1318 fixed
sibling_count4 -0.08506 0.06395 -1.33 -0.2104 0.04028 fixed
sibling_count5 -0.1773 0.0723 -2.452 -0.319 -0.03557 fixed
sibling_count5+ -0.3553 0.07036 -5.05 -0.4932 -0.2174 fixed
birth_order_nonlinear2 0.01388 0.04185 0.3317 -0.06814 0.0959 fixed
birth_order_nonlinear3 0.05777 0.04965 1.164 -0.03954 0.1551 fixed
birth_order_nonlinear4 0.1137 0.06092 1.867 -0.005655 0.2332 fixed
birth_order_nonlinear5 0.07464 0.07384 1.011 -0.07009 0.2194 fixed
birth_order_nonlinear5+ 0.1402 0.07456 1.88 -0.005951 0.2863 fixed
sd_(Intercept).mother_pidlink 0.5365 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8329 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -4.353 4.016 -1.084 -12.22 3.518 fixed
poly(age, 3, raw = TRUE)1 0.5416 0.4178 1.296 -0.2772 1.36 fixed
poly(age, 3, raw = TRUE)2 -0.01934 0.01428 -1.354 -0.04733 0.00865 fixed
poly(age, 3, raw = TRUE)3 0.0002189 0.0001605 1.364 -0.00009562 0.0005334 fixed
male -0.1694 0.03192 -5.307 -0.2319 -0.1068 fixed
count_birth_order2/2 -0.1002 0.07959 -1.259 -0.2562 0.0558 fixed
count_birth_order1/3 -0.02078 0.07425 -0.2798 -0.1663 0.1247 fixed
count_birth_order2/3 -0.03022 0.08127 -0.3718 -0.1895 0.1291 fixed
count_birth_order3/3 0.04444 0.09002 0.4936 -0.132 0.2209 fixed
count_birth_order1/4 -0.1305 0.08648 -1.509 -0.3 0.03902 fixed
count_birth_order2/4 -0.08433 0.08829 -0.9552 -0.2574 0.08871 fixed
count_birth_order3/4 -0.1842 0.08944 -2.06 -0.3595 -0.008904 fixed
count_birth_order4/4 0.09782 0.09722 1.006 -0.09272 0.2884 fixed
count_birth_order1/5 -0.2027 0.1116 -1.817 -0.4214 0.016 fixed
count_birth_order2/5 -0.2693 0.1155 -2.331 -0.4957 -0.04286 fixed
count_birth_order3/5 -0.06099 0.1052 -0.5797 -0.2672 0.1452 fixed
count_birth_order4/5 -0.143 0.1047 -1.366 -0.3482 0.0622 fixed
count_birth_order5/5 -0.1681 0.1096 -1.534 -0.3829 0.04671 fixed
count_birth_order1/5+ -0.5594 0.1064 -5.259 -0.7679 -0.3509 fixed
count_birth_order2/5+ -0.2231 0.1032 -2.162 -0.4254 -0.02086 fixed
count_birth_order3/5+ -0.2829 0.09714 -2.912 -0.4733 -0.09253 fixed
count_birth_order4/5+ -0.3577 0.09658 -3.704 -0.547 -0.1684 fixed
count_birth_order5/5+ -0.3057 0.09391 -3.255 -0.4898 -0.1217 fixed
count_birth_order5+/5+ -0.257 0.07455 -3.447 -0.4031 -0.1109 fixed
sd_(Intercept).mother_pidlink 0.5403 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8298 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 9998 10066 -4988 9976 NA NA NA
12 9996 10071 -4986 9972 3.826 1 0.05046
16 10003 10102 -4985 9971 1.676 4 0.7951
26 10004 10165 -4976 9952 18.3 10 0.05009

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -4.371 4.006 -1.091 -12.22 3.481 fixed
poly(age, 3, raw = TRUE)1 0.542 0.4168 1.3 -0.275 1.359 fixed
poly(age, 3, raw = TRUE)2 -0.01939 0.01425 -1.361 -0.04731 0.008538 fixed
poly(age, 3, raw = TRUE)3 0.0002186 0.0001601 1.365 -0.0000952 0.0005324 fixed
male -0.1633 0.03187 -5.124 -0.2258 -0.1008 fixed
sibling_count3 0.0165 0.06515 0.2532 -0.1112 0.1442 fixed
sibling_count4 -0.03107 0.06602 -0.4706 -0.1605 0.09832 fixed
sibling_count5 -0.05786 0.06883 -0.8406 -0.1928 0.07704 fixed
sibling_count5+ -0.1946 0.05987 -3.25 -0.3119 -0.07722 fixed
sd_(Intercept).mother_pidlink 0.5427 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8324 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -4.479 4.007 -1.118 -12.33 3.375 fixed
birth_order 0.008948 0.009208 0.9718 -0.009099 0.027 fixed
poly(age, 3, raw = TRUE)1 0.5518 0.417 1.323 -0.2654 1.369 fixed
poly(age, 3, raw = TRUE)2 -0.01974 0.01425 -1.385 -0.04767 0.008196 fixed
poly(age, 3, raw = TRUE)3 0.000223 0.0001602 1.392 -0.00009093 0.0005369 fixed
male -0.1641 0.03188 -5.147 -0.2266 -0.1016 fixed
sibling_count3 0.01225 0.06531 0.1876 -0.1157 0.1402 fixed
sibling_count4 -0.04063 0.06675 -0.6087 -0.1715 0.09019 fixed
sibling_count5 -0.073 0.07058 -1.034 -0.2113 0.06533 fixed
sibling_count5+ -0.2265 0.06829 -3.316 -0.3603 -0.09263 fixed
sd_(Intercept).mother_pidlink 0.543 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8323 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -4.65 4.016 -1.158 -12.52 3.221 fixed
poly(age, 3, raw = TRUE)1 0.5695 0.4178 1.363 -0.2494 1.388 fixed
poly(age, 3, raw = TRUE)2 -0.02033 0.01428 -1.424 -0.04832 0.007657 fixed
poly(age, 3, raw = TRUE)3 0.0002296 0.0001605 1.431 -0.0000849 0.0005441 fixed
male -0.1645 0.0319 -5.156 -0.227 -0.102 fixed
sibling_count3 0.00755 0.0661 0.1142 -0.122 0.1371 fixed
sibling_count4 -0.05017 0.06842 -0.7333 -0.1843 0.08393 fixed
sibling_count5 -0.07514 0.07309 -1.028 -0.2184 0.06812 fixed
sibling_count5+ -0.2374 0.06988 -3.398 -0.3744 -0.1005 fixed
birth_order_nonlinear2 0.02522 0.04284 0.5888 -0.05874 0.1092 fixed
birth_order_nonlinear3 0.04141 0.04964 0.8342 -0.05589 0.1387 fixed
birth_order_nonlinear4 0.05611 0.05947 0.9435 -0.06045 0.1727 fixed
birth_order_nonlinear5 0.003816 0.07089 0.05383 -0.1351 0.1428 fixed
birth_order_nonlinear5+ 0.09122 0.06817 1.338 -0.04239 0.2248 fixed
sd_(Intercept).mother_pidlink 0.5431 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8326 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -5.133 4.028 -1.274 -13.03 2.762 fixed
poly(age, 3, raw = TRUE)1 0.6228 0.4191 1.486 -0.1986 1.444 fixed
poly(age, 3, raw = TRUE)2 -0.02212 0.01432 -1.544 -0.05019 0.005953 fixed
poly(age, 3, raw = TRUE)3 0.0002494 0.0001609 1.55 -0.00006603 0.0005648 fixed
male -0.1678 0.03197 -5.248 -0.2304 -0.1051 fixed
count_birth_order2/2 -0.08389 0.09032 -0.9288 -0.2609 0.09314 fixed
count_birth_order1/3 -0.05503 0.08281 -0.6645 -0.2173 0.1073 fixed
count_birth_order2/3 -0.004886 0.08999 -0.05429 -0.1813 0.1715 fixed
count_birth_order3/3 0.05245 0.09768 0.5369 -0.139 0.2439 fixed
count_birth_order1/4 -0.09204 0.0918 -1.003 -0.272 0.08788 fixed
count_birth_order2/4 -0.02352 0.09274 -0.2537 -0.2053 0.1582 fixed
count_birth_order3/4 -0.1511 0.0978 -1.546 -0.3428 0.04053 fixed
count_birth_order4/4 0.03295 0.1078 0.3057 -0.1783 0.2443 fixed
count_birth_order1/5 -0.09179 0.1055 -0.8701 -0.2985 0.115 fixed
count_birth_order2/5 -0.09788 0.1095 -0.8938 -0.3125 0.1168 fixed
count_birth_order3/5 -0.127 0.1091 -1.164 -0.3409 0.08683 fixed
count_birth_order4/5 -0.1385 0.1098 -1.262 -0.3537 0.07666 fixed
count_birth_order5/5 0.007127 0.1138 0.06262 -0.2159 0.2302 fixed
count_birth_order1/5+ -0.3253 0.09756 -3.334 -0.5165 -0.1341 fixed
count_birth_order2/5+ -0.2307 0.09925 -2.325 -0.4252 -0.03618 fixed
count_birth_order3/5+ -0.1457 0.09446 -1.542 -0.3308 0.03948 fixed
count_birth_order4/5+ -0.2183 0.09352 -2.334 -0.4016 -0.03499 fixed
count_birth_order5/5+ -0.3437 0.09551 -3.599 -0.5309 -0.1565 fixed
count_birth_order5+/5+ -0.1877 0.07701 -2.437 -0.3387 -0.03677 fixed
sd_(Intercept).mother_pidlink 0.5425 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8327 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 10073 10141 -5026 10051 NA NA NA
12 10074 10149 -5025 10050 0.9464 1 0.3306
16 10081 10180 -5024 10049 1.653 4 0.7992
26 10089 10250 -5019 10037 11.21 10 0.3414

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -4.667 4.039 -1.156 -12.58 3.249 fixed
poly(age, 3, raw = TRUE)1 0.5704 0.42 1.358 -0.2529 1.394 fixed
poly(age, 3, raw = TRUE)2 -0.02031 0.01435 -1.415 -0.04844 0.007817 fixed
poly(age, 3, raw = TRUE)3 0.0002285 0.0001612 1.418 -0.00008744 0.0005445 fixed
male -0.172 0.03217 -5.347 -0.235 -0.1089 fixed
sibling_count3 0.04253 0.05789 0.7347 -0.07093 0.156 fixed
sibling_count4 -0.00362 0.06007 -0.06026 -0.1213 0.1141 fixed
sibling_count5 -0.08955 0.06729 -1.331 -0.2214 0.04234 fixed
sibling_count5+ -0.2283 0.05746 -3.972 -0.3409 -0.1156 fixed
sd_(Intercept).mother_pidlink 0.5385 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8325 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -4.89 4.039 -1.211 -12.81 3.027 fixed
birth_order 0.01969 0.01069 1.842 -0.001265 0.04065 fixed
poly(age, 3, raw = TRUE)1 0.5898 0.42 1.404 -0.2335 1.413 fixed
poly(age, 3, raw = TRUE)2 -0.02099 0.01435 -1.462 -0.04912 0.00714 fixed
poly(age, 3, raw = TRUE)3 0.0002369 0.0001612 1.47 -0.00007906 0.0005529 fixed
male -0.1735 0.03217 -5.394 -0.2366 -0.1105 fixed
sibling_count3 0.03345 0.05807 0.576 -0.08038 0.1473 fixed
sibling_count4 -0.02628 0.06129 -0.4288 -0.1464 0.09385 fixed
sibling_count5 -0.125 0.06996 -1.786 -0.2621 0.01215 fixed
sibling_count5+ -0.2998 0.06933 -4.324 -0.4357 -0.1639 fixed
sd_(Intercept).mother_pidlink 0.5383 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8323 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -4.722 4.046 -1.167 -12.65 3.208 fixed
poly(age, 3, raw = TRUE)1 0.5745 0.4208 1.365 -0.2502 1.399 fixed
poly(age, 3, raw = TRUE)2 -0.02048 0.01438 -1.424 -0.04865 0.007704 fixed
poly(age, 3, raw = TRUE)3 0.0002313 0.0001615 1.432 -0.0000852 0.0005478 fixed
male -0.173 0.03218 -5.378 -0.2361 -0.11 fixed
sibling_count3 0.03542 0.05889 0.6013 -0.08001 0.1508 fixed
sibling_count4 -0.03158 0.06339 -0.4983 -0.1558 0.09266 fixed
sibling_count5 -0.1308 0.07281 -1.797 -0.2735 0.0119 fixed
sibling_count5+ -0.3123 0.07097 -4.4 -0.4514 -0.1732 fixed
birth_order_nonlinear2 0.01662 0.04175 0.3981 -0.06521 0.09844 fixed
birth_order_nonlinear3 0.02904 0.04951 0.5866 -0.068 0.1261 fixed
birth_order_nonlinear4 0.09316 0.06189 1.505 -0.02814 0.2145 fixed
birth_order_nonlinear5 0.07755 0.07635 1.016 -0.0721 0.2272 fixed
birth_order_nonlinear5+ 0.1508 0.07683 1.962 0.000189 0.3014 fixed
sd_(Intercept).mother_pidlink 0.5379 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8329 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -4.752 4.051 -1.173 -12.69 3.188 fixed
poly(age, 3, raw = TRUE)1 0.5826 0.4213 1.383 -0.2431 1.408 fixed
poly(age, 3, raw = TRUE)2 -0.02079 0.01439 -1.445 -0.04901 0.007415 fixed
poly(age, 3, raw = TRUE)3 0.0002354 0.0001617 1.456 -0.00008143 0.0005523 fixed
male -0.1735 0.03224 -5.382 -0.2367 -0.1103 fixed
count_birth_order2/2 -0.08293 0.07738 -1.072 -0.2346 0.06873 fixed
count_birth_order1/3 -0.004316 0.07332 -0.05886 -0.148 0.1394 fixed
count_birth_order2/3 0.0333 0.08109 0.4107 -0.1256 0.1922 fixed
count_birth_order3/3 0.01099 0.0878 0.1252 -0.1611 0.1831 fixed
count_birth_order1/4 -0.05723 0.08623 -0.6637 -0.2262 0.1118 fixed
count_birth_order2/4 -0.02736 0.088 -0.3109 -0.1998 0.1451 fixed
count_birth_order3/4 -0.1187 0.08945 -1.327 -0.294 0.05662 fixed
count_birth_order4/4 0.06962 0.09595 0.7256 -0.1184 0.2577 fixed
count_birth_order1/5 -0.1861 0.1109 -1.678 -0.4035 0.03122 fixed
count_birth_order2/5 -0.2571 0.1174 -2.19 -0.4873 -0.02698 fixed
count_birth_order3/5 -0.0593 0.1067 -0.5557 -0.2684 0.1498 fixed
count_birth_order4/5 -0.08314 0.1076 -0.7725 -0.2941 0.1278 fixed
count_birth_order5/5 -0.04927 0.1164 -0.4234 -0.2773 0.1788 fixed
count_birth_order1/5+ -0.4789 0.109 -4.393 -0.6926 -0.2652 fixed
count_birth_order2/5+ -0.208 0.1068 -1.946 -0.4174 0.001458 fixed
count_birth_order3/5+ -0.2551 0.0975 -2.616 -0.4462 -0.06399 fixed
count_birth_order4/5+ -0.2976 0.09981 -2.982 -0.4932 -0.102 fixed
count_birth_order5/5+ -0.2926 0.09505 -3.078 -0.4789 -0.1063 fixed
count_birth_order5+/5+ -0.1995 0.07533 -2.648 -0.3472 -0.05186 fixed
sd_(Intercept).mother_pidlink 0.5395 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8319 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 9830 9898 -4904 9808 NA NA NA
12 9829 9903 -4903 9805 3.4 1 0.06519
16 9836 9934 -4902 9804 1.376 4 0.8484
26 9845 10005 -4896 9793 11 10 0.3572

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Raven 2007 young

birthorder <- birthorder %>% mutate(outcome = raven_2007_young)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
             data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.538 0.4647 -14.07 -7.449 -5.628 fixed
poly(age, 3, raw = TRUE)1 0.7174 0.06304 11.38 0.5939 0.841 fixed
poly(age, 3, raw = TRUE)2 -0.02329 0.00279 -8.346 -0.02876 -0.01782 fixed
poly(age, 3, raw = TRUE)3 0.0002269 0.00004068 5.579 0.0001472 0.0003067 fixed
male 0.04906 0.01897 2.586 0.01188 0.08624 fixed
sibling_count3 -0.02344 0.04132 -0.5672 -0.1044 0.05754 fixed
sibling_count4 -0.1274 0.04218 -3.02 -0.21 -0.04471 fixed
sibling_count5 -0.1396 0.04548 -3.07 -0.2288 -0.05048 fixed
sibling_count5+ -0.3188 0.03645 -8.748 -0.3903 -0.2474 fixed
sd_(Intercept).mother_pidlink 0.5465 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7892 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.539 0.4647 -14.07 -7.45 -5.628 fixed
birth_order -0.0008314 0.004314 -0.1927 -0.009287 0.007625 fixed
poly(age, 3, raw = TRUE)1 0.7178 0.06307 11.38 0.5942 0.8414 fixed
poly(age, 3, raw = TRUE)2 -0.02331 0.002792 -8.348 -0.02878 -0.01783 fixed
poly(age, 3, raw = TRUE)3 0.0002271 0.00004069 5.582 0.0001474 0.0003069 fixed
male 0.04913 0.01897 2.589 0.01194 0.08631 fixed
sibling_count3 -0.02302 0.04138 -0.5563 -0.1041 0.05808 fixed
sibling_count4 -0.1264 0.04246 -2.978 -0.2096 -0.04322 fixed
sibling_count5 -0.1381 0.0462 -2.988 -0.2286 -0.04751 fixed
sibling_count5+ -0.3144 0.04308 -7.298 -0.3988 -0.23 fixed
sd_(Intercept).mother_pidlink 0.5466 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7893 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.59 0.4654 -14.16 -7.502 -5.678 fixed
poly(age, 3, raw = TRUE)1 0.7218 0.06309 11.44 0.5981 0.8454 fixed
poly(age, 3, raw = TRUE)2 -0.02346 0.002792 -8.403 -0.02894 -0.01799 fixed
poly(age, 3, raw = TRUE)3 0.0002291 0.00004069 5.63 0.0001494 0.0003089 fixed
male 0.04929 0.01897 2.598 0.01211 0.08648 fixed
sibling_count3 -0.02145 0.04194 -0.5114 -0.1036 0.06075 fixed
sibling_count4 -0.1281 0.04391 -2.917 -0.2142 -0.04204 fixed
sibling_count5 -0.1345 0.04857 -2.769 -0.2297 -0.03927 fixed
sibling_count5+ -0.3092 0.04591 -6.733 -0.3991 -0.2192 fixed
birth_order_nonlinear2 0.06258 0.02815 2.223 0.007396 0.1178 fixed
birth_order_nonlinear3 -0.01065 0.03274 -0.3254 -0.07482 0.05352 fixed
birth_order_nonlinear4 0.02709 0.03839 0.7056 -0.04816 0.1023 fixed
birth_order_nonlinear5 -0.001689 0.04332 -0.039 -0.08659 0.08321 fixed
birth_order_nonlinear5+ 0.006358 0.03898 0.1631 -0.07005 0.08277 fixed
sd_(Intercept).mother_pidlink 0.5461 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7893 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.638 0.4657 -14.25 -7.55 -5.725 fixed
poly(age, 3, raw = TRUE)1 0.7279 0.06307 11.54 0.6043 0.8515 fixed
poly(age, 3, raw = TRUE)2 -0.0237 0.002791 -8.491 -0.02917 -0.01823 fixed
poly(age, 3, raw = TRUE)3 0.0002319 0.00004068 5.701 0.0001522 0.0003116 fixed
male 0.04812 0.01897 2.537 0.01094 0.08529 fixed
count_birth_order2/2 0.06072 0.05725 1.061 -0.05149 0.1729 fixed
count_birth_order1/3 -0.0343 0.05105 -0.6719 -0.1344 0.06576 fixed
count_birth_order2/3 0.07454 0.05626 1.325 -0.03572 0.1848 fixed
count_birth_order3/3 -0.05386 0.06313 -0.8531 -0.1776 0.06987 fixed
count_birth_order1/4 -0.1607 0.05803 -2.769 -0.2744 -0.04693 fixed
count_birth_order2/4 -0.07732 0.05939 -1.302 -0.1937 0.03909 fixed
count_birth_order3/4 -0.1763 0.06208 -2.84 -0.298 -0.05465 fixed
count_birth_order4/4 0.01525 0.06777 0.2251 -0.1176 0.1481 fixed
count_birth_order1/5 -0.1478 0.07044 -2.098 -0.2858 -0.009719 fixed
count_birth_order2/5 -0.09451 0.07441 -1.27 -0.2403 0.05133 fixed
count_birth_order3/5 -0.2273 0.07096 -3.204 -0.3664 -0.08824 fixed
count_birth_order4/5 -0.1158 0.07449 -1.554 -0.2618 0.03024 fixed
count_birth_order5/5 -0.02131 0.07249 -0.294 -0.1634 0.1208 fixed
count_birth_order1/5+ -0.2317 0.06517 -3.555 -0.3594 -0.1039 fixed
count_birth_order2/5+ -0.271 0.06388 -4.242 -0.3962 -0.1458 fixed
count_birth_order3/5+ -0.2168 0.05985 -3.622 -0.3341 -0.09946 fixed
count_birth_order4/5+ -0.3418 0.05625 -6.076 -0.452 -0.2315 fixed
count_birth_order5/5+ -0.3567 0.05421 -6.58 -0.4629 -0.2504 fixed
count_birth_order5+/5+ -0.3052 0.0421 -7.248 -0.3877 -0.2226 fixed
sd_(Intercept).mother_pidlink 0.5453 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7889 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 24804 24882 -12391 24782 NA NA NA
12 24806 24891 -12391 24782 0.03716 1 0.8471
16 24807 24921 -12387 24775 7.146 4 0.1284
26 24803 24988 -12376 24751 23.59 10 0.008753

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.861 0.9731 -7.051 -8.769 -4.954 fixed
poly(age, 3, raw = TRUE)1 0.7568 0.1455 5.202 0.4717 1.042 fixed
poly(age, 3, raw = TRUE)2 -0.02486 0.00708 -3.511 -0.03873 -0.01098 fixed
poly(age, 3, raw = TRUE)3 0.0002526 0.0001124 2.248 0.00003234 0.0004728 fixed
male 0.01685 0.0235 0.7172 -0.0292 0.0629 fixed
sibling_count3 -0.01718 0.03598 -0.4774 -0.0877 0.05334 fixed
sibling_count4 -0.1219 0.04074 -2.991 -0.2017 -0.04201 fixed
sibling_count5 -0.1907 0.04921 -3.875 -0.2871 -0.09422 fixed
sibling_count5+ -0.2719 0.04397 -6.184 -0.3581 -0.1857 fixed
sd_(Intercept).mother_pidlink 0.5182 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7306 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.915 0.9742 -7.098 -8.824 -5.005 fixed
birth_order 0.009749 0.009035 1.079 -0.007959 0.02746 fixed
poly(age, 3, raw = TRUE)1 0.7625 0.1456 5.238 0.4772 1.048 fixed
poly(age, 3, raw = TRUE)2 -0.02513 0.007083 -3.548 -0.03901 -0.01125 fixed
poly(age, 3, raw = TRUE)3 0.0002572 0.0001124 2.288 0.00003685 0.0004776 fixed
male 0.01654 0.0235 0.7041 -0.02951 0.06259 fixed
sibling_count3 -0.02201 0.03626 -0.6069 -0.09308 0.04907 fixed
sibling_count4 -0.1335 0.04217 -3.167 -0.2162 -0.0509 fixed
sibling_count5 -0.2107 0.0526 -4.005 -0.3138 -0.1076 fixed
sibling_count5+ -0.315 0.05944 -5.3 -0.4315 -0.1986 fixed
sd_(Intercept).mother_pidlink 0.5188 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7303 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.964 0.9744 -7.147 -8.874 -5.054 fixed
poly(age, 3, raw = TRUE)1 0.7685 0.1456 5.279 0.4831 1.054 fixed
poly(age, 3, raw = TRUE)2 -0.02539 0.007084 -3.584 -0.03927 -0.01151 fixed
poly(age, 3, raw = TRUE)3 0.000261 0.0001124 2.322 0.00004065 0.0004813 fixed
male 0.01645 0.0235 0.7003 -0.0296 0.0625 fixed
sibling_count3 -0.02539 0.03711 -0.6843 -0.09813 0.04734 fixed
sibling_count4 -0.1275 0.04426 -2.88 -0.2142 -0.04071 fixed
sibling_count5 -0.2088 0.05603 -3.726 -0.3186 -0.09897 fixed
sibling_count5+ -0.3376 0.06226 -5.423 -0.4596 -0.2156 fixed
birth_order_nonlinear2 0.05199 0.02881 1.805 -0.004469 0.1085 fixed
birth_order_nonlinear3 0.03421 0.03777 0.9058 -0.03981 0.1082 fixed
birth_order_nonlinear4 0.001927 0.04906 0.03928 -0.09422 0.09808 fixed
birth_order_nonlinear5 0.07045 0.06369 1.106 -0.05438 0.1953 fixed
birth_order_nonlinear5+ 0.1232 0.06923 1.78 -0.01244 0.2589 fixed
sd_(Intercept).mother_pidlink 0.5199 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7295 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.96 0.9751 -7.138 -8.871 -5.049 fixed
poly(age, 3, raw = TRUE)1 0.7688 0.1457 5.276 0.4832 1.054 fixed
poly(age, 3, raw = TRUE)2 -0.0254 0.007091 -3.582 -0.0393 -0.0115 fixed
poly(age, 3, raw = TRUE)3 0.0002612 0.0001125 2.321 0.00004067 0.0004818 fixed
male 0.01598 0.02351 0.6795 -0.03011 0.06206 fixed
count_birth_order2/2 0.02934 0.04729 0.6205 -0.06334 0.122 fixed
count_birth_order1/3 -0.05929 0.04542 -1.305 -0.1483 0.02973 fixed
count_birth_order2/3 0.04985 0.04937 1.01 -0.04693 0.1466 fixed
count_birth_order3/3 0.01367 0.05506 0.2483 -0.09425 0.1216 fixed
count_birth_order1/4 -0.1267 0.06105 -2.076 -0.2464 -0.007086 fixed
count_birth_order2/4 -0.1538 0.06097 -2.523 -0.2733 -0.03431 fixed
count_birth_order3/4 -0.1289 0.06357 -2.028 -0.2535 -0.004344 fixed
count_birth_order4/4 -0.04741 0.06315 -0.7507 -0.1712 0.07637 fixed
count_birth_order1/5 -0.1687 0.09218 -1.83 -0.3493 0.01201 fixed
count_birth_order2/5 -0.1148 0.09718 -1.181 -0.3052 0.07568 fixed
count_birth_order3/5 -0.2755 0.08563 -3.218 -0.4434 -0.1077 fixed
count_birth_order4/5 -0.2379 0.08049 -2.956 -0.3957 -0.08014 fixed
count_birth_order5/5 -0.1169 0.07802 -1.499 -0.2698 0.03599 fixed
count_birth_order1/5+ -0.2922 0.1129 -2.589 -0.5134 -0.07097 fixed
count_birth_order2/5+ -0.2078 0.1108 -1.876 -0.4248 0.009291 fixed
count_birth_order3/5+ -0.1477 0.09608 -1.537 -0.336 0.04059 fixed
count_birth_order4/5+ -0.4902 0.08614 -5.691 -0.659 -0.3214 fixed
count_birth_order5/5+ -0.3071 0.07735 -3.971 -0.4587 -0.1555 fixed
count_birth_order5+/5+ -0.2259 0.05481 -4.121 -0.3333 -0.1184 fixed
sd_(Intercept).mother_pidlink 0.5188 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7294 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 13765 13838 -6872 13743 NA NA NA
12 13766 13845 -6871 13742 1.164 1 0.2806
16 13769 13874 -6868 13737 5.381 4 0.2504
26 13771 13942 -6860 13719 17.62 10 0.06172

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.83 0.9707 -7.036 -8.733 -4.927 fixed
poly(age, 3, raw = TRUE)1 0.7549 0.1451 5.203 0.4705 1.039 fixed
poly(age, 3, raw = TRUE)2 -0.0248 0.007062 -3.511 -0.03864 -0.01096 fixed
poly(age, 3, raw = TRUE)3 0.000252 0.0001121 2.248 0.00003226 0.0004717 fixed
male 0.0151 0.0234 0.6455 -0.03075 0.06096 fixed
sibling_count3 -0.03337 0.03864 -0.8637 -0.1091 0.04236 fixed
sibling_count4 -0.08849 0.0421 -2.102 -0.171 -0.00597 fixed
sibling_count5 -0.1713 0.04757 -3.6 -0.2645 -0.07803 fixed
sibling_count5+ -0.2459 0.04111 -5.981 -0.3265 -0.1653 fixed
sd_(Intercept).mother_pidlink 0.5178 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.73 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.846 0.9721 -7.042 -8.751 -4.94 fixed
birth_order 0.002372 0.007711 0.3077 -0.01274 0.01749 fixed
poly(age, 3, raw = TRUE)1 0.7566 0.1452 5.21 0.472 1.041 fixed
poly(age, 3, raw = TRUE)2 -0.02488 0.007068 -3.521 -0.03873 -0.01103 fixed
poly(age, 3, raw = TRUE)3 0.0002533 0.0001122 2.258 0.00003344 0.0004732 fixed
male 0.015 0.0234 0.6411 -0.03086 0.06087 fixed
sibling_count3 -0.03458 0.03885 -0.8902 -0.1107 0.04156 fixed
sibling_count4 -0.09117 0.043 -2.12 -0.1755 -0.006886 fixed
sibling_count5 -0.1758 0.04978 -3.531 -0.2733 -0.0782 fixed
sibling_count5+ -0.2556 0.05184 -4.931 -0.3572 -0.154 fixed
sd_(Intercept).mother_pidlink 0.518 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7299 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.917 0.9721 -7.115 -8.822 -5.011 fixed
poly(age, 3, raw = TRUE)1 0.763 0.1452 5.254 0.4784 1.048 fixed
poly(age, 3, raw = TRUE)2 -0.02513 0.007068 -3.556 -0.03898 -0.01128 fixed
poly(age, 3, raw = TRUE)3 0.0002564 0.0001122 2.286 0.00003657 0.0004763 fixed
male 0.01531 0.0234 0.6541 -0.03056 0.06117 fixed
sibling_count3 -0.03284 0.03966 -0.828 -0.1106 0.04489 fixed
sibling_count4 -0.08939 0.04488 -1.992 -0.1774 -0.001434 fixed
sibling_count5 -0.1838 0.0527 -3.487 -0.2871 -0.08049 fixed
sibling_count5+ -0.2573 0.05375 -4.787 -0.3626 -0.1519 fixed
birth_order_nonlinear2 0.07217 0.02953 2.444 0.0143 0.13 fixed
birth_order_nonlinear3 -0.005236 0.03733 -0.1403 -0.07839 0.06792 fixed
birth_order_nonlinear4 0.02286 0.04685 0.4879 -0.06896 0.1147 fixed
birth_order_nonlinear5 0.07559 0.05817 1.299 -0.03842 0.1896 fixed
birth_order_nonlinear5+ 0.033 0.05905 0.5588 -0.08273 0.1487 fixed
sd_(Intercept).mother_pidlink 0.5181 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7295 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.899 0.9721 -7.097 -8.804 -4.994 fixed
poly(age, 3, raw = TRUE)1 0.7606 0.1453 5.236 0.4759 1.045 fixed
poly(age, 3, raw = TRUE)2 -0.02507 0.007069 -3.546 -0.03892 -0.01121 fixed
poly(age, 3, raw = TRUE)3 0.0002566 0.0001122 2.287 0.00003666 0.0004765 fixed
male 0.01386 0.02341 0.5922 -0.03201 0.05974 fixed
count_birth_order2/2 0.08613 0.05191 1.659 -0.01561 0.1879 fixed
count_birth_order1/3 -0.06727 0.04895 -1.374 -0.1632 0.02867 fixed
count_birth_order2/3 0.05816 0.05259 1.106 -0.04492 0.1612 fixed
count_birth_order3/3 0.02034 0.05905 0.3444 -0.0954 0.1361 fixed
count_birth_order1/4 -0.07108 0.06131 -1.159 -0.1912 0.04909 fixed
count_birth_order2/4 -0.06081 0.06189 -0.9826 -0.1821 0.06049 fixed
count_birth_order3/4 -0.1245 0.06503 -1.914 -0.2519 0.002971 fixed
count_birth_order4/4 0.005626 0.06667 0.08438 -0.125 0.1363 fixed
count_birth_order1/5 -0.07135 0.08316 -0.858 -0.2343 0.09164 fixed
count_birth_order2/5 -0.1042 0.0841 -1.239 -0.269 0.06065 fixed
count_birth_order3/5 -0.3061 0.08024 -3.815 -0.4633 -0.1488 fixed
count_birth_order4/5 -0.1987 0.08071 -2.461 -0.3569 -0.04048 fixed
count_birth_order5/5 -0.06386 0.07738 -0.8252 -0.2155 0.08781 fixed
count_birth_order1/5+ -0.2017 0.08248 -2.446 -0.3634 -0.04007 fixed
count_birth_order2/5+ -0.1627 0.09223 -1.764 -0.3435 0.01808 fixed
count_birth_order3/5+ -0.1926 0.08051 -2.392 -0.3504 -0.03479 fixed
count_birth_order4/5+ -0.2925 0.07496 -3.902 -0.4394 -0.1456 fixed
count_birth_order5/5+ -0.2174 0.07194 -3.022 -0.3584 -0.07641 fixed
count_birth_order5+/5+ -0.2229 0.05227 -4.265 -0.3254 -0.1205 fixed
sd_(Intercept).mother_pidlink 0.5195 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7283 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 13877 13950 -6928 13855 NA NA NA
12 13879 13958 -6927 13855 0.09453 1 0.7585
16 13878 13984 -6923 13846 8.425 4 0.07721
26 13884 14056 -6916 13832 14.31 10 0.1592

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.615 0.9829 -6.73 -8.542 -4.689 fixed
poly(age, 3, raw = TRUE)1 0.7211 0.1469 4.91 0.4333 1.009 fixed
poly(age, 3, raw = TRUE)2 -0.02326 0.007145 -3.256 -0.03726 -0.009257 fixed
poly(age, 3, raw = TRUE)3 0.000229 0.0001133 2.02 0.000006808 0.0004511 fixed
male 0.01755 0.02385 0.736 -0.02919 0.0643 fixed
sibling_count3 -0.01117 0.03603 -0.31 -0.08178 0.05944 fixed
sibling_count4 -0.1152 0.04097 -2.812 -0.1955 -0.03491 fixed
sibling_count5 -0.1965 0.05121 -3.837 -0.2968 -0.0961 fixed
sibling_count5+ -0.2427 0.0452 -5.369 -0.3313 -0.1541 fixed
sd_(Intercept).mother_pidlink 0.5228 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.731 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.68 0.9842 -6.787 -8.609 -4.751 fixed
birth_order 0.01105 0.009284 1.19 -0.007151 0.02924 fixed
poly(age, 3, raw = TRUE)1 0.7283 0.147 4.955 0.4402 1.016 fixed
poly(age, 3, raw = TRUE)2 -0.02361 0.007149 -3.303 -0.03762 -0.009598 fixed
poly(age, 3, raw = TRUE)3 0.0002349 0.0001134 2.071 0.00001261 0.0004572 fixed
male 0.01736 0.02385 0.7279 -0.02938 0.0641 fixed
sibling_count3 -0.01649 0.03631 -0.4542 -0.08766 0.05468 fixed
sibling_count4 -0.1285 0.04248 -3.026 -0.2118 -0.04528 fixed
sibling_count5 -0.2186 0.05449 -4.011 -0.3254 -0.1118 fixed
sibling_count5+ -0.2912 0.06091 -4.781 -0.4106 -0.1718 fixed
sd_(Intercept).mother_pidlink 0.5235 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7306 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.689 0.9848 -6.792 -8.619 -4.759 fixed
poly(age, 3, raw = TRUE)1 0.7288 0.147 4.956 0.4406 1.017 fixed
poly(age, 3, raw = TRUE)2 -0.02361 0.007152 -3.302 -0.03763 -0.009594 fixed
poly(age, 3, raw = TRUE)3 0.0002347 0.0001135 2.069 0.00001234 0.0004571 fixed
male 0.01744 0.02385 0.731 -0.02931 0.06419 fixed
sibling_count3 -0.02629 0.03714 -0.7078 -0.09908 0.04651 fixed
sibling_count4 -0.1254 0.04463 -2.81 -0.2129 -0.03793 fixed
sibling_count5 -0.2056 0.05789 -3.552 -0.3191 -0.09215 fixed
sibling_count5+ -0.302 0.0639 -4.727 -0.4273 -0.1768 fixed
birth_order_nonlinear2 0.05142 0.02888 1.78 -0.005198 0.108 fixed
birth_order_nonlinear3 0.06668 0.03813 1.749 -0.008058 0.1414 fixed
birth_order_nonlinear4 -0.007045 0.05027 -0.1401 -0.1056 0.09148 fixed
birth_order_nonlinear5 0.02412 0.06693 0.3604 -0.1071 0.1553 fixed
birth_order_nonlinear5+ 0.1217 0.07164 1.699 -0.01869 0.2621 fixed
sd_(Intercept).mother_pidlink 0.5233 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7304 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.726 0.9854 -6.825 -8.657 -4.795 fixed
poly(age, 3, raw = TRUE)1 0.7345 0.1472 4.991 0.4461 1.023 fixed
poly(age, 3, raw = TRUE)2 -0.02385 0.007159 -3.332 -0.03788 -0.009823 fixed
poly(age, 3, raw = TRUE)3 0.0002381 0.0001136 2.097 0.00001553 0.0004607 fixed
male 0.01742 0.02388 0.7295 -0.02938 0.06422 fixed
count_birth_order2/2 0.02951 0.04629 0.6375 -0.06121 0.1202 fixed
count_birth_order1/3 -0.06344 0.04539 -1.398 -0.1524 0.02553 fixed
count_birth_order2/3 0.06221 0.04994 1.246 -0.03567 0.1601 fixed
count_birth_order3/3 0.03591 0.05522 0.6502 -0.07233 0.1441 fixed
count_birth_order1/4 -0.1158 0.062 -1.867 -0.2373 0.00576 fixed
count_birth_order2/4 -0.1576 0.06146 -2.564 -0.278 -0.03711 fixed
count_birth_order3/4 -0.1036 0.06379 -1.624 -0.2286 0.02142 fixed
count_birth_order4/4 -0.05084 0.06338 -0.8021 -0.1751 0.07338 fixed
count_birth_order1/5 -0.1952 0.0937 -2.083 -0.3788 -0.01153 fixed
count_birth_order2/5 -0.1358 0.1035 -1.312 -0.3386 0.06706 fixed
count_birth_order3/5 -0.189 0.09007 -2.098 -0.3656 -0.01248 fixed
count_birth_order4/5 -0.2471 0.08507 -2.905 -0.4139 -0.08038 fixed
count_birth_order5/5 -0.1613 0.08299 -1.943 -0.3239 0.001389 fixed
count_birth_order1/5+ -0.1999 0.1156 -1.729 -0.4266 0.02673 fixed
count_birth_order2/5+ -0.204 0.1144 -1.783 -0.4282 0.02024 fixed
count_birth_order3/5+ -0.08644 0.09791 -0.8828 -0.2783 0.1055 fixed
count_birth_order4/5+ -0.4909 0.09081 -5.405 -0.6689 -0.3129 fixed
count_birth_order5/5+ -0.3151 0.08087 -3.897 -0.4736 -0.1566 fixed
count_birth_order5+/5+ -0.1932 0.05652 -3.418 -0.304 -0.08242 fixed
sd_(Intercept).mother_pidlink 0.5227 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.73 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 13447 13519 -6713 13425 NA NA NA
12 13448 13526 -6712 13424 1.415 1 0.2342
16 13449 13554 -6708 13417 6.748 4 0.1498
26 13451 13621 -6699 13399 18.11 10 0.05321

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Numeracy 2007 young

birthorder <- birthorder %>% mutate(outcome = math_2007_young)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
             data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.754 0.4653 -14.52 -7.666 -5.842 fixed
poly(age, 3, raw = TRUE)1 0.7929 0.06319 12.55 0.6691 0.9168 fixed
poly(age, 3, raw = TRUE)2 -0.02804 0.002801 -10.01 -0.03352 -0.02255 fixed
poly(age, 3, raw = TRUE)3 0.0003055 0.00004089 7.472 0.0002254 0.0003856 fixed
male -0.1437 0.01893 -7.589 -0.1808 -0.1066 fixed
sibling_count3 0.0287 0.04191 0.6847 -0.05345 0.1108 fixed
sibling_count4 -0.1166 0.04282 -2.724 -0.2006 -0.03272 fixed
sibling_count5 -0.0236 0.04622 -0.5106 -0.1142 0.06699 fixed
sibling_count5+ -0.2097 0.037 -5.668 -0.2822 -0.1372 fixed
sd_(Intercept).mother_pidlink 0.5785 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7773 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.752 0.4653 -14.51 -7.664 -5.84 fixed
birth_order 0.001749 0.004367 0.4004 -0.00681 0.01031 fixed
poly(age, 3, raw = TRUE)1 0.7921 0.06322 12.53 0.6682 0.916 fixed
poly(age, 3, raw = TRUE)2 -0.028 0.002802 -9.99 -0.03349 -0.0225 fixed
poly(age, 3, raw = TRUE)3 0.000305 0.0000409 7.457 0.0002249 0.0003852 fixed
male -0.1438 0.01893 -7.595 -0.1809 -0.1067 fixed
sibling_count3 0.02781 0.04197 0.6625 -0.05446 0.1101 fixed
sibling_count4 -0.1186 0.04311 -2.752 -0.2031 -0.03414 fixed
sibling_count5 -0.02692 0.04696 -0.5733 -0.119 0.06512 fixed
sibling_count5+ -0.2191 0.04378 -5.005 -0.3049 -0.1333 fixed
sd_(Intercept).mother_pidlink 0.5787 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7773 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.789 0.4661 -14.57 -7.702 -5.875 fixed
poly(age, 3, raw = TRUE)1 0.7953 0.06325 12.57 0.6713 0.9193 fixed
poly(age, 3, raw = TRUE)2 -0.02812 0.002803 -10.03 -0.03362 -0.02263 fixed
poly(age, 3, raw = TRUE)3 0.0003066 0.00004091 7.494 0.0002264 0.0003868 fixed
male -0.1439 0.01893 -7.602 -0.181 -0.1068 fixed
sibling_count3 0.02289 0.04252 0.5382 -0.06046 0.1062 fixed
sibling_count4 -0.1276 0.04453 -2.865 -0.2149 -0.04033 fixed
sibling_count5 -0.02192 0.04929 -0.4446 -0.1185 0.0747 fixed
sibling_count5+ -0.218 0.04654 -4.684 -0.3092 -0.1268 fixed
birth_order_nonlinear2 0.04666 0.02802 1.665 -0.008259 0.1016 fixed
birth_order_nonlinear3 0.02498 0.0326 0.7662 -0.03892 0.08888 fixed
birth_order_nonlinear4 0.0334 0.03827 0.8727 -0.0416 0.1084 fixed
birth_order_nonlinear5 -0.03609 0.0432 -0.8353 -0.1208 0.04859 fixed
birth_order_nonlinear5+ 0.031 0.03917 0.7913 -0.04577 0.1078 fixed
sd_(Intercept).mother_pidlink 0.5785 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7773 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.828 0.4667 -14.63 -7.743 -5.913 fixed
poly(age, 3, raw = TRUE)1 0.7989 0.06328 12.62 0.6749 0.9229 fixed
poly(age, 3, raw = TRUE)2 -0.02826 0.002804 -10.08 -0.03376 -0.02277 fixed
poly(age, 3, raw = TRUE)3 0.0003083 0.00004092 7.535 0.0002281 0.0003886 fixed
male -0.144 0.01894 -7.6 -0.1811 -0.1068 fixed
count_birth_order2/2 0.08251 0.0571 1.445 -0.0294 0.1944 fixed
count_birth_order1/3 0.04288 0.05146 0.8333 -0.05798 0.1437 fixed
count_birth_order2/3 0.08563 0.05661 1.513 -0.02533 0.1966 fixed
count_birth_order3/3 0.03055 0.06347 0.4813 -0.09384 0.1549 fixed
count_birth_order1/4 -0.1306 0.05842 -2.235 -0.2451 -0.01608 fixed
count_birth_order2/4 -0.08114 0.05974 -1.358 -0.1982 0.03595 fixed
count_birth_order3/4 -0.1099 0.06239 -1.761 -0.2322 0.01243 fixed
count_birth_order4/4 -0.02188 0.06808 -0.3214 -0.1553 0.1116 fixed
count_birth_order1/5 -0.062 0.0708 -0.8757 -0.2008 0.07676 fixed
count_birth_order2/5 -0.01654 0.07469 -0.2215 -0.1629 0.1299 fixed
count_birth_order3/5 0.07044 0.07124 0.9888 -0.06918 0.2101 fixed
count_birth_order4/5 0.06037 0.07477 0.8075 -0.08617 0.2069 fixed
count_birth_order5/5 -0.03592 0.07285 -0.4931 -0.1787 0.1069 fixed
count_birth_order1/5+ -0.1415 0.06539 -2.164 -0.2697 -0.01333 fixed
count_birth_order2/5+ -0.1494 0.06406 -2.332 -0.2749 -0.02384 fixed
count_birth_order3/5+ -0.1759 0.06005 -2.929 -0.2936 -0.05817 fixed
count_birth_order4/5+ -0.2217 0.05646 -3.927 -0.3324 -0.1111 fixed
count_birth_order5/5+ -0.2462 0.05442 -4.524 -0.3529 -0.1396 fixed
count_birth_order5+/5+ -0.1776 0.04252 -4.176 -0.2609 -0.09422 fixed
sd_(Intercept).mother_pidlink 0.5778 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7777 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 24867 24945 -12423 24845 NA NA NA
12 24869 24954 -12422 24845 0.1601 1 0.6891
16 24871 24985 -12419 24839 6.098 4 0.1919
26 24882 25068 -12415 24830 8.388 10 0.591

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -7.329 1.011 -7.253 -9.31 -5.349 fixed
poly(age, 3, raw = TRUE)1 0.862 0.1511 5.705 0.5659 1.158 fixed
poly(age, 3, raw = TRUE)2 -0.03069 0.007352 -4.174 -0.0451 -0.01628 fixed
poly(age, 3, raw = TRUE)3 0.0003444 0.0001167 2.952 0.0001157 0.0005731 fixed
male -0.1627 0.02442 -6.663 -0.2106 -0.1148 fixed
sibling_count3 0.03514 0.03749 0.9373 -0.03834 0.1086 fixed
sibling_count4 -0.03986 0.04246 -0.9387 -0.1231 0.04336 fixed
sibling_count5 -0.09433 0.05129 -1.839 -0.1949 0.006192 fixed
sibling_count5+ -0.1596 0.04583 -3.483 -0.2494 -0.0698 fixed
sd_(Intercept).mother_pidlink 0.5442 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7571 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -7.448 1.011 -7.364 -9.431 -5.466 fixed
birth_order 0.02171 0.009399 2.31 0.003291 0.04013 fixed
poly(age, 3, raw = TRUE)1 0.8745 0.1511 5.787 0.5783 1.171 fixed
poly(age, 3, raw = TRUE)2 -0.03129 0.007353 -4.255 -0.0457 -0.01687 fixed
poly(age, 3, raw = TRUE)3 0.0003545 0.0001167 3.038 0.0001258 0.0005833 fixed
male -0.1634 0.02441 -6.693 -0.2112 -0.1155 fixed
sibling_count3 0.02437 0.03776 0.6453 -0.04965 0.09839 fixed
sibling_count4 -0.06593 0.04392 -1.501 -0.152 0.02015 fixed
sibling_count5 -0.139 0.0548 -2.537 -0.2464 -0.0316 fixed
sibling_count5+ -0.2558 0.06191 -4.132 -0.3771 -0.1345 fixed
sd_(Intercept).mother_pidlink 0.5443 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7566 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -7.45 1.013 -7.356 -9.435 -5.465 fixed
poly(age, 3, raw = TRUE)1 0.8771 0.1513 5.797 0.5806 1.174 fixed
poly(age, 3, raw = TRUE)2 -0.03139 0.007362 -4.264 -0.04582 -0.01696 fixed
poly(age, 3, raw = TRUE)3 0.0003557 0.0001168 3.044 0.0001267 0.0005847 fixed
male -0.1636 0.02443 -6.696 -0.2115 -0.1157 fixed
sibling_count3 0.02301 0.03864 0.5955 -0.05272 0.09875 fixed
sibling_count4 -0.07166 0.04609 -1.555 -0.162 0.01868 fixed
sibling_count5 -0.1373 0.05835 -2.353 -0.2516 -0.02293 fixed
sibling_count5+ -0.2439 0.06483 -3.762 -0.371 -0.1168 fixed
birth_order_nonlinear2 0.03194 0.02994 1.067 -0.02674 0.09062 fixed
birth_order_nonlinear3 0.0497 0.03926 1.266 -0.02725 0.1266 fixed
birth_order_nonlinear4 0.08489 0.05101 1.664 -0.01509 0.1849 fixed
birth_order_nonlinear5 0.06655 0.06622 1.005 -0.06324 0.1963 fixed
birth_order_nonlinear5+ 0.1277 0.07203 1.774 -0.01342 0.2689 fixed
sd_(Intercept).mother_pidlink 0.5439 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7573 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -7.405 1.014 -7.305 -9.392 -5.419 fixed
poly(age, 3, raw = TRUE)1 0.8706 0.1515 5.747 0.5737 1.168 fixed
poly(age, 3, raw = TRUE)2 -0.03108 0.007371 -4.216 -0.04553 -0.01663 fixed
poly(age, 3, raw = TRUE)3 0.0003509 0.000117 3 0.0001216 0.0005802 fixed
male -0.1619 0.02446 -6.622 -0.2099 -0.114 fixed
count_birth_order2/2 0.0297 0.04917 0.6042 -0.06666 0.1261 fixed
count_birth_order1/3 0.03687 0.04727 0.7801 -0.05577 0.1295 fixed
count_birth_order2/3 0.05894 0.05138 1.147 -0.04177 0.1596 fixed
count_birth_order3/3 0.03767 0.0573 0.6574 -0.07463 0.15 fixed
count_birth_order1/4 -0.07716 0.06353 -1.215 -0.2017 0.04735 fixed
count_birth_order2/4 -0.05468 0.06344 -0.862 -0.179 0.06965 fixed
count_birth_order3/4 -0.02054 0.06614 -0.3106 -0.1502 0.1091 fixed
count_birth_order4/4 0.02997 0.06571 0.456 -0.09883 0.1588 fixed
count_birth_order1/5 -0.09647 0.09591 -1.006 -0.2844 0.09151 fixed
count_birth_order2/5 -0.2193 0.1011 -2.17 -0.4174 -0.02121 fixed
count_birth_order3/5 -0.1128 0.08908 -1.266 -0.2874 0.06179 fixed
count_birth_order4/5 -0.09586 0.08374 -1.145 -0.26 0.06827 fixed
count_birth_order5/5 0.01419 0.08118 0.1748 -0.1449 0.1733 fixed
count_birth_order1/5+ -0.4417 0.1174 -3.762 -0.6718 -0.2116 fixed
count_birth_order2/5+ -0.04547 0.1152 -0.3947 -0.2712 0.1803 fixed
count_birth_order3/5+ -0.05332 0.09993 -0.5336 -0.2492 0.1425 fixed
count_birth_order4/5+ -0.1371 0.08959 -1.531 -0.3127 0.03845 fixed
count_birth_order5/5+ -0.2583 0.08046 -3.21 -0.416 -0.1006 fixed
count_birth_order5+/5+ -0.1188 0.05706 -2.082 -0.2306 -0.006936 fixed
sd_(Intercept).mother_pidlink 0.543 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7573 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 14186 14258 -7082 14164 NA NA NA
12 14183 14262 -7079 14159 5.343 1 0.0208
16 14191 14297 -7080 14159 0 4 1
26 14196 14368 -7072 14144 14.79 10 0.1397

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -7.308 1.01 -7.239 -9.287 -5.329 fixed
poly(age, 3, raw = TRUE)1 0.8603 0.1509 5.701 0.5645 1.156 fixed
poly(age, 3, raw = TRUE)2 -0.03059 0.007344 -4.166 -0.04499 -0.0162 fixed
poly(age, 3, raw = TRUE)3 0.0003423 0.0001166 2.937 0.0001138 0.0005708 fixed
male -0.1626 0.02436 -6.674 -0.2103 -0.1148 fixed
sibling_count3 0.01646 0.04035 0.4081 -0.06262 0.09554 fixed
sibling_count4 -0.04713 0.04397 -1.072 -0.1333 0.03906 fixed
sibling_count5 -0.06369 0.04968 -1.282 -0.1611 0.03369 fixed
sibling_count5+ -0.1238 0.04294 -2.883 -0.2079 -0.03963 fixed
sd_(Intercept).mother_pidlink 0.5459 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7572 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -7.388 1.011 -7.309 -9.369 -5.407 fixed
birth_order 0.01223 0.00804 1.521 -0.00353 0.02799 fixed
poly(age, 3, raw = TRUE)1 0.8691 0.151 5.757 0.5732 1.165 fixed
poly(age, 3, raw = TRUE)2 -0.03101 0.007348 -4.22 -0.04541 -0.01661 fixed
poly(age, 3, raw = TRUE)3 0.0003492 0.0001166 2.994 0.0001206 0.0005778 fixed
male -0.1631 0.02436 -6.695 -0.2108 -0.1153 fixed
sibling_count3 0.01024 0.04055 0.2526 -0.06924 0.08973 fixed
sibling_count4 -0.06099 0.04491 -1.358 -0.149 0.02702 fixed
sibling_count5 -0.08694 0.05198 -1.673 -0.1888 0.01494 fixed
sibling_count5+ -0.1739 0.05412 -3.213 -0.28 -0.06782 fixed
sd_(Intercept).mother_pidlink 0.5461 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.757 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -7.397 1.012 -7.31 -9.38 -5.414 fixed
poly(age, 3, raw = TRUE)1 0.8706 0.1512 5.76 0.5744 1.167 fixed
poly(age, 3, raw = TRUE)2 -0.03105 0.007356 -4.221 -0.04547 -0.01663 fixed
poly(age, 3, raw = TRUE)3 0.0003493 0.0001168 2.992 0.0001205 0.0005781 fixed
male -0.1633 0.02438 -6.697 -0.211 -0.1155 fixed
sibling_count3 0.00672 0.04141 0.1623 -0.07445 0.08789 fixed
sibling_count4 -0.06608 0.04687 -1.41 -0.1579 0.02578 fixed
sibling_count5 -0.08536 0.05504 -1.551 -0.1932 0.02252 fixed
sibling_count5+ -0.1641 0.05613 -2.924 -0.2741 -0.05413 fixed
birth_order_nonlinear2 0.03599 0.03073 1.171 -0.02424 0.09623 fixed
birth_order_nonlinear3 0.03925 0.03887 1.01 -0.03693 0.1154 fixed
birth_order_nonlinear4 0.05036 0.0488 1.032 -0.04528 0.146 fixed
birth_order_nonlinear5 0.03331 0.06059 0.5497 -0.08546 0.1521 fixed
birth_order_nonlinear5+ 0.07217 0.06157 1.172 -0.04851 0.1928 fixed
sd_(Intercept).mother_pidlink 0.5455 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7577 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -7.423 1.013 -7.324 -9.409 -5.436 fixed
poly(age, 3, raw = TRUE)1 0.8736 0.1514 5.77 0.5768 1.17 fixed
poly(age, 3, raw = TRUE)2 -0.0312 0.007368 -4.235 -0.04565 -0.01676 fixed
poly(age, 3, raw = TRUE)3 0.0003521 0.000117 3.01 0.0001228 0.0005813 fixed
male -0.1627 0.02441 -6.664 -0.2105 -0.1148 fixed
count_birth_order2/2 0.05236 0.05412 0.9676 -0.0537 0.1584 fixed
count_birth_order1/3 0.01647 0.05107 0.3225 -0.08363 0.1166 fixed
count_birth_order2/3 0.05293 0.05487 0.9647 -0.05461 0.1605 fixed
count_birth_order3/3 0.03427 0.0616 0.5564 -0.08646 0.155 fixed
count_birth_order1/4 -0.1059 0.06396 -1.656 -0.2312 0.01946 fixed
count_birth_order2/4 -0.04341 0.06455 -0.6724 -0.1699 0.08311 fixed
count_birth_order3/4 -0.03786 0.06782 -0.5582 -0.1708 0.09507 fixed
count_birth_order4/4 0.0801 0.06955 1.152 -0.05621 0.2164 fixed
count_birth_order1/5 -0.04695 0.08674 -0.5413 -0.217 0.1231 fixed
count_birth_order2/5 -0.05659 0.08771 -0.6452 -0.2285 0.1153 fixed
count_birth_order3/5 0.01862 0.08368 0.2225 -0.1454 0.1826 fixed
count_birth_order4/5 -0.1552 0.08418 -1.844 -0.3202 0.009763 fixed
count_birth_order5/5 -0.0149 0.08072 -0.1846 -0.1731 0.1433 fixed
count_birth_order1/5+ -0.09441 0.08602 -1.097 -0.263 0.07419 fixed
count_birth_order2/5+ -0.1298 0.09618 -1.35 -0.3183 0.05868 fixed
count_birth_order3/5+ -0.1148 0.08396 -1.367 -0.2794 0.04975 fixed
count_birth_order4/5+ -0.1252 0.07818 -1.602 -0.2784 0.02802 fixed
count_birth_order5/5+ -0.1599 0.07502 -2.131 -0.3069 -0.01284 fixed
count_birth_order5+/5+ -0.0894 0.05454 -1.639 -0.1963 0.01751 fixed
sd_(Intercept).mother_pidlink 0.5441 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7586 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 14321 14393 -7149 14299 NA NA NA
12 14320 14399 -7148 14296 2.316 1 0.128
16 14328 14434 -7148 14296 0.2387 4 0.9934
26 14339 14511 -7144 14287 8.723 10 0.5586

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -7.342 1.021 -7.189 -9.344 -5.341 fixed
poly(age, 3, raw = TRUE)1 0.8591 0.1526 5.63 0.56 1.158 fixed
poly(age, 3, raw = TRUE)2 -0.03046 0.007422 -4.103 -0.045 -0.01591 fixed
poly(age, 3, raw = TRUE)3 0.0003395 0.0001178 2.883 0.0001087 0.0005703 fixed
male -0.1635 0.02479 -6.595 -0.2121 -0.1149 fixed
sibling_count3 0.06614 0.03751 1.763 -0.007369 0.1397 fixed
sibling_count4 -0.01357 0.04265 -0.3182 -0.09718 0.07003 fixed
sibling_count5 -0.1193 0.05332 -2.238 -0.2238 -0.0148 fixed
sibling_count5+ -0.1277 0.04707 -2.714 -0.22 -0.03547 fixed
sd_(Intercept).mother_pidlink 0.5469 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7586 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -7.466 1.022 -7.304 -9.47 -5.463 fixed
birth_order 0.02146 0.009655 2.223 0.002538 0.04038 fixed
poly(age, 3, raw = TRUE)1 0.8728 0.1526 5.718 0.5736 1.172 fixed
poly(age, 3, raw = TRUE)2 -0.03112 0.007425 -4.192 -0.04568 -0.01657 fixed
poly(age, 3, raw = TRUE)3 0.0003509 0.0001178 2.979 0.00012 0.0005818 fixed
male -0.1639 0.02478 -6.612 -0.2124 -0.1153 fixed
sibling_count3 0.0558 0.03778 1.477 -0.01826 0.1299 fixed
sibling_count4 -0.03953 0.04421 -0.8941 -0.1262 0.04712 fixed
sibling_count5 -0.1623 0.05671 -2.863 -0.2735 -0.05119 fixed
sibling_count5+ -0.2221 0.06338 -3.504 -0.3463 -0.09788 fixed
sd_(Intercept).mother_pidlink 0.547 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7581 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -7.462 1.024 -7.289 -9.469 -5.456 fixed
poly(age, 3, raw = TRUE)1 0.8742 0.1529 5.719 0.5746 1.174 fixed
poly(age, 3, raw = TRUE)2 -0.03116 0.007435 -4.191 -0.04573 -0.01659 fixed
poly(age, 3, raw = TRUE)3 0.0003508 0.000118 2.974 0.0001196 0.000582 fixed
male -0.1639 0.02481 -6.606 -0.2125 -0.1152 fixed
sibling_count3 0.05035 0.03866 1.302 -0.02543 0.1261 fixed
sibling_count4 -0.04989 0.04646 -1.074 -0.141 0.04117 fixed
sibling_count5 -0.1631 0.06026 -2.707 -0.2812 -0.04499 fixed
sibling_count5+ -0.2083 0.06652 -3.131 -0.3386 -0.07789 fixed
birth_order_nonlinear2 0.03225 0.03003 1.074 -0.0266 0.09111 fixed
birth_order_nonlinear3 0.06694 0.03965 1.688 -0.01077 0.1446 fixed
birth_order_nonlinear4 0.08688 0.05228 1.662 -0.01558 0.1893 fixed
birth_order_nonlinear5 0.0601 0.0696 0.8634 -0.07632 0.1965 fixed
birth_order_nonlinear5+ 0.1203 0.07453 1.614 -0.02578 0.2664 fixed
sd_(Intercept).mother_pidlink 0.5465 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7587 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -7.458 1.025 -7.275 -9.467 -5.449 fixed
poly(age, 3, raw = TRUE)1 0.8732 0.1531 5.703 0.5731 1.173 fixed
poly(age, 3, raw = TRUE)2 -0.03111 0.007447 -4.178 -0.04571 -0.01652 fixed
poly(age, 3, raw = TRUE)3 0.0003503 0.0001181 2.965 0.0001187 0.0005818 fixed
male -0.1617 0.02485 -6.51 -0.2104 -0.113 fixed
count_birth_order2/2 0.04034 0.04816 0.8377 -0.05404 0.1347 fixed
count_birth_order1/3 0.06742 0.04725 1.427 -0.02518 0.16 fixed
count_birth_order2/3 0.09286 0.05198 1.787 -0.009014 0.1947 fixed
count_birth_order3/3 0.08153 0.05748 1.418 -0.03112 0.1942 fixed
count_birth_order1/4 -0.07282 0.06453 -1.128 -0.1993 0.05366 fixed
count_birth_order2/4 -0.02447 0.06396 -0.3826 -0.1498 0.1009 fixed
count_birth_order3/4 0.03036 0.06638 0.4573 -0.09975 0.1605 fixed
count_birth_order4/4 0.0658 0.06596 0.9975 -0.06348 0.1951 fixed
count_birth_order1/5 -0.05713 0.09751 -0.5859 -0.2483 0.134 fixed
count_birth_order2/5 -0.2773 0.1077 -2.576 -0.4884 -0.06629 fixed
count_birth_order3/5 -0.109 0.09373 -1.163 -0.2928 0.07467 fixed
count_birth_order4/5 -0.1463 0.08853 -1.652 -0.3198 0.02724 fixed
count_birth_order5/5 -0.02612 0.08637 -0.3025 -0.1954 0.1432 fixed
count_birth_order1/5+ -0.3727 0.1203 -3.097 -0.6086 -0.1369 fixed
count_birth_order2/5+ -0.0671 0.119 -0.5637 -0.3004 0.1662 fixed
count_birth_order3/5+ -0.01418 0.1019 -0.1391 -0.2139 0.1855 fixed
count_birth_order4/5+ -0.08858 0.09449 -0.9374 -0.2738 0.09663 fixed
count_birth_order5/5+ -0.2156 0.08415 -2.562 -0.3805 -0.05066 fixed
count_birth_order5+/5+ -0.08683 0.05884 -1.476 -0.2022 0.02849 fixed
sd_(Intercept).mother_pidlink 0.5454 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.759 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 13855 13928 -6917 13833 NA NA NA
12 13852 13931 -6914 13828 4.947 1 0.02613
16 13860 13965 -6914 13828 0.1435 4 0.9975
26 13867 14037 -6907 13815 13.41 10 0.2014

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Counting backwards

birthorder <- birthorder %>% mutate(outcome = count_backwards)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5158 0.1653 -3.12 -0.8398 -0.1918 fixed
poly(age, 3, raw = TRUE)1 0.06972 0.01599 4.359 0.03837 0.1011 fixed
poly(age, 3, raw = TRUE)2 -0.002356 0.0004765 -4.944 -0.00329 -0.001422 fixed
poly(age, 3, raw = TRUE)3 0.00002179 0.000004446 4.9 0.00001307 0.0000305 fixed
male 0.04321 0.01663 2.599 0.01062 0.0758 fixed
sibling_count3 0.002307 0.03444 0.06697 -0.0652 0.06981 fixed
sibling_count4 0.03135 0.03566 0.8792 -0.03854 0.1012 fixed
sibling_count5 0.0143 0.03708 0.3855 -0.05838 0.08698 fixed
sibling_count5+ -0.1004 0.0291 -3.449 -0.1574 -0.04334 fixed
sd_(Intercept).mother_pidlink 0.4186 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8969 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.521 0.1653 -3.152 -0.845 -0.1971 fixed
birth_order -0.007541 0.003559 -2.119 -0.01452 -0.0005663 fixed
poly(age, 3, raw = TRUE)1 0.07236 0.01604 4.511 0.04092 0.1038 fixed
poly(age, 3, raw = TRUE)2 -0.002453 0.0004786 -5.125 -0.003391 -0.001515 fixed
poly(age, 3, raw = TRUE)3 0.0000227 0.000004466 5.083 0.00001395 0.00003145 fixed
male 0.04349 0.01663 2.615 0.01089 0.07608 fixed
sibling_count3 0.00394 0.03443 0.1144 -0.06355 0.07143 fixed
sibling_count4 0.03655 0.03572 1.023 -0.03346 0.1066 fixed
sibling_count5 0.0234 0.03731 0.6272 -0.04972 0.09652 fixed
sibling_count5+ -0.07194 0.03204 -2.246 -0.1347 -0.009149 fixed
sd_(Intercept).mother_pidlink 0.4172 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8973 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5287 0.1657 -3.19 -0.8536 -0.2039 fixed
poly(age, 3, raw = TRUE)1 0.07327 0.01604 4.567 0.04183 0.1047 fixed
poly(age, 3, raw = TRUE)2 -0.002467 0.0004786 -5.154 -0.003405 -0.001529 fixed
poly(age, 3, raw = TRUE)3 0.00002265 0.000004468 5.069 0.00001389 0.0000314 fixed
male 0.04386 0.01662 2.638 0.01128 0.07644 fixed
sibling_count3 0.0181 0.03489 0.5187 -0.05029 0.08648 fixed
sibling_count4 0.05936 0.03668 1.618 -0.01253 0.1313 fixed
sibling_count5 0.05141 0.03865 1.33 -0.02434 0.1272 fixed
sibling_count5+ -0.04218 0.03354 -1.258 -0.1079 0.02356 fixed
birth_order_nonlinear2 -0.03426 0.02401 -1.427 -0.08133 0.0128 fixed
birth_order_nonlinear3 -0.08943 0.02829 -3.162 -0.1449 -0.03399 fixed
birth_order_nonlinear4 -0.08064 0.03218 -2.506 -0.1437 -0.01756 fixed
birth_order_nonlinear5 -0.08633 0.03661 -2.358 -0.1581 -0.01458 fixed
birth_order_nonlinear5+ -0.09508 0.03072 -3.095 -0.1553 -0.03486 fixed
sd_(Intercept).mother_pidlink 0.4188 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8965 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5266 0.1665 -3.163 -0.8529 -0.2003 fixed
poly(age, 3, raw = TRUE)1 0.07317 0.01605 4.56 0.04172 0.1046 fixed
poly(age, 3, raw = TRUE)2 -0.002458 0.0004789 -5.132 -0.003396 -0.001519 fixed
poly(age, 3, raw = TRUE)3 0.00002251 0.000004471 5.034 0.00001374 0.00003127 fixed
male 0.04425 0.01663 2.661 0.01166 0.07683 fixed
count_birth_order2/2 -0.04338 0.04687 -0.9254 -0.1352 0.04849 fixed
count_birth_order1/3 0.03772 0.04498 0.8386 -0.05044 0.1259 fixed
count_birth_order2/3 -0.0197 0.05012 -0.393 -0.1179 0.07853 fixed
count_birth_order3/3 -0.1232 0.05612 -2.195 -0.2332 -0.01317 fixed
count_birth_order1/4 0.0815 0.05147 1.583 -0.01938 0.1824 fixed
count_birth_order2/4 0.005026 0.05402 0.09305 -0.1009 0.1109 fixed
count_birth_order3/4 -0.06449 0.05836 -1.105 -0.1789 0.0499 fixed
count_birth_order4/4 -0.009001 0.06157 -0.1462 -0.1297 0.1117 fixed
count_birth_order1/5 -0.0384 0.05782 -0.6642 -0.1517 0.07492 fixed
count_birth_order2/5 0.06285 0.06083 1.033 -0.05637 0.1821 fixed
count_birth_order3/5 0.03161 0.06263 0.5048 -0.09113 0.1544 fixed
count_birth_order4/5 -0.05632 0.06619 -0.851 -0.186 0.0734 fixed
count_birth_order5/5 -0.03531 0.06734 -0.5243 -0.1673 0.09668 fixed
count_birth_order1/5+ -0.05063 0.04693 -1.079 -0.1426 0.04136 fixed
count_birth_order2/5+ -0.08895 0.04831 -1.841 -0.1836 0.005739 fixed
count_birth_order3/5+ -0.1235 0.04744 -2.604 -0.2165 -0.03054 fixed
count_birth_order4/5+ -0.1254 0.04636 -2.705 -0.2163 -0.03454 fixed
count_birth_order5/5+ -0.1327 0.0467 -2.842 -0.2243 -0.04119 fixed
count_birth_order5+/5+ -0.1408 0.03701 -3.804 -0.2133 -0.06823 fixed
sd_(Intercept).mother_pidlink 0.4189 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8965 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 38110 38193 -19044 38088 NA NA NA
12 38108 38198 -19042 38084 4.489 1 0.03411
16 38105 38225 -19037 38073 10.57 4 0.03183
26 38116 38311 -19032 38064 9.266 10 0.507

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.742 0.4065 -4.284 -2.538 -0.9448 fixed
poly(age, 3, raw = TRUE)1 0.2094 0.04611 4.543 0.1191 0.2998 fixed
poly(age, 3, raw = TRUE)2 -0.006882 0.001645 -4.185 -0.01011 -0.003659 fixed
poly(age, 3, raw = TRUE)3 0.00007036 0.00001862 3.779 0.00003387 0.0001069 fixed
male -0.002035 0.02375 -0.08571 -0.04858 0.04451 fixed
sibling_count3 -0.00724 0.0379 -0.191 -0.08152 0.06704 fixed
sibling_count4 -0.0264 0.04103 -0.6435 -0.1068 0.05401 fixed
sibling_count5 -0.07347 0.04681 -1.569 -0.1652 0.01828 fixed
sibling_count5+ -0.1963 0.04127 -4.756 -0.2772 -0.1154 fixed
sd_(Intercept).mother_pidlink 0.3845 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8347 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.734 0.4066 -4.266 -2.531 -0.9374 fixed
birth_order -0.008164 0.00792 -1.031 -0.02369 0.00736 fixed
poly(age, 3, raw = TRUE)1 0.2098 0.04611 4.55 0.1194 0.3002 fixed
poly(age, 3, raw = TRUE)2 -0.006883 0.001645 -4.185 -0.01011 -0.003659 fixed
poly(age, 3, raw = TRUE)3 0.00007005 0.00001862 3.762 0.00003355 0.0001065 fixed
male -0.001657 0.02375 -0.06975 -0.04821 0.04489 fixed
sibling_count3 -0.003215 0.03809 -0.08441 -0.07787 0.07144 fixed
sibling_count4 -0.01684 0.04205 -0.4004 -0.09925 0.06558 fixed
sibling_count5 -0.05794 0.04916 -1.179 -0.1543 0.03842 fixed
sibling_count5+ -0.1653 0.05105 -3.238 -0.2653 -0.06523 fixed
sd_(Intercept).mother_pidlink 0.3836 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.835 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.751 0.4073 -4.3 -2.549 -0.953 fixed
poly(age, 3, raw = TRUE)1 0.2103 0.04614 4.558 0.1199 0.3008 fixed
poly(age, 3, raw = TRUE)2 -0.006901 0.001646 -4.193 -0.01013 -0.003675 fixed
poly(age, 3, raw = TRUE)3 0.00007025 0.00001864 3.769 0.00003372 0.0001068 fixed
male -0.001182 0.02375 -0.04974 -0.04774 0.04538 fixed
sibling_count3 0.007384 0.03885 0.1901 -0.06877 0.08353 fixed
sibling_count4 -0.003047 0.04364 -0.06983 -0.08858 0.08249 fixed
sibling_count5 -0.04944 0.05155 -0.9591 -0.1505 0.05159 fixed
sibling_count5+ -0.1575 0.05246 -3.002 -0.2603 -0.05469 fixed
birth_order_nonlinear2 0.003951 0.03053 0.1294 -0.05588 0.06379 fixed
birth_order_nonlinear3 -0.06261 0.03762 -1.664 -0.1363 0.01111 fixed
birth_order_nonlinear4 -0.03636 0.0466 -0.7802 -0.1277 0.05497 fixed
birth_order_nonlinear5 -0.007188 0.05817 -0.1236 -0.1212 0.1068 fixed
birth_order_nonlinear5+ -0.05932 0.05891 -1.007 -0.1748 0.05615 fixed
sd_(Intercept).mother_pidlink 0.3846 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8348 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.724 0.408 -4.227 -2.524 -0.9248 fixed
poly(age, 3, raw = TRUE)1 0.205 0.04622 4.435 0.1144 0.2956 fixed
poly(age, 3, raw = TRUE)2 -0.0067 0.001649 -4.063 -0.009932 -0.003468 fixed
poly(age, 3, raw = TRUE)3 0.00006785 0.00001868 3.633 0.00003125 0.0001045 fixed
male -0.003637 0.02376 -0.1531 -0.05021 0.04294 fixed
count_birth_order2/2 0.06035 0.0554 1.089 -0.04823 0.1689 fixed
count_birth_order1/3 0.04137 0.04935 0.8382 -0.05536 0.1381 fixed
count_birth_order2/3 0.006316 0.05378 0.1175 -0.09909 0.1117 fixed
count_birth_order3/3 -0.03376 0.05998 -0.5628 -0.1513 0.08381 fixed
count_birth_order1/4 -0.03262 0.06047 -0.5395 -0.1511 0.08589 fixed
count_birth_order2/4 0.0837 0.06257 1.338 -0.03892 0.2063 fixed
count_birth_order3/4 -0.07312 0.06565 -1.114 -0.2018 0.05555 fixed
count_birth_order4/4 -0.002899 0.06845 -0.04236 -0.1371 0.1313 fixed
count_birth_order1/5 0.01161 0.08139 0.1427 -0.1479 0.1711 fixed
count_birth_order2/5 -0.07328 0.08756 -0.837 -0.2449 0.09833 fixed
count_birth_order3/5 -0.1484 0.0822 -1.806 -0.3096 0.01267 fixed
count_birth_order4/5 -0.03242 0.07917 -0.4095 -0.1876 0.1228 fixed
count_birth_order5/5 -0.02795 0.08174 -0.342 -0.1882 0.1322 fixed
count_birth_order1/5+ -0.05036 0.0809 -0.6225 -0.2089 0.1082 fixed
count_birth_order2/5+ -0.2429 0.0813 -2.987 -0.4022 -0.08351 fixed
count_birth_order3/5+ -0.1141 0.08049 -1.418 -0.2719 0.04363 fixed
count_birth_order4/5+ -0.227 0.07547 -3.008 -0.3749 -0.07911 fixed
count_birth_order5/5+ -0.1526 0.07184 -2.124 -0.2934 -0.01176 fixed
count_birth_order5+/5+ -0.1988 0.05463 -3.639 -0.3059 -0.09173 fixed
sd_(Intercept).mother_pidlink 0.385 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8344 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 15526 15600 -7752 15504 NA NA NA
12 15527 15607 -7752 15503 1.066 1 0.3018
16 15532 15639 -7750 15500 3.189 4 0.5268
26 15540 15713 -7744 15488 12.37 10 0.2613

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.735 0.4053 -4.282 -2.53 -0.9409 fixed
poly(age, 3, raw = TRUE)1 0.2105 0.04599 4.578 0.1204 0.3006 fixed
poly(age, 3, raw = TRUE)2 -0.006937 0.001641 -4.228 -0.01015 -0.003721 fixed
poly(age, 3, raw = TRUE)3 0.00007079 0.00001858 3.81 0.00003437 0.0001072 fixed
male -0.001301 0.02365 -0.05501 -0.04766 0.04506 fixed
sibling_count3 -0.03781 0.04097 -0.923 -0.1181 0.04248 fixed
sibling_count4 -0.007388 0.04336 -0.1704 -0.09238 0.0776 fixed
sibling_count5 -0.06853 0.04634 -1.479 -0.1594 0.02229 fixed
sibling_count5+ -0.1464 0.04069 -3.598 -0.2261 -0.06664 fixed
sd_(Intercept).mother_pidlink 0.3853 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8346 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.721 0.4053 -4.247 -2.516 -0.927 fixed
birth_order -0.01171 0.006915 -1.694 -0.02527 0.001839 fixed
poly(age, 3, raw = TRUE)1 0.2106 0.04598 4.58 0.1205 0.3007 fixed
poly(age, 3, raw = TRUE)2 -0.006923 0.001641 -4.22 -0.01014 -0.003707 fixed
poly(age, 3, raw = TRUE)3 0.00007016 0.00001858 3.776 0.00003374 0.0001066 fixed
male -0.0008451 0.02365 -0.03573 -0.0472 0.04551 fixed
sibling_count3 -0.03205 0.04109 -0.7799 -0.1126 0.04849 fixed
sibling_count4 0.005742 0.04403 0.1304 -0.08055 0.09203 fixed
sibling_count5 -0.04796 0.04788 -1.002 -0.1418 0.04588 fixed
sibling_count5+ -0.1033 0.04796 -2.154 -0.1973 -0.009283 fixed
sd_(Intercept).mother_pidlink 0.384 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8349 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.737 0.4058 -4.28 -2.532 -0.9414 fixed
poly(age, 3, raw = TRUE)1 0.2108 0.046 4.584 0.1207 0.301 fixed
poly(age, 3, raw = TRUE)2 -0.006929 0.001641 -4.222 -0.01015 -0.003712 fixed
poly(age, 3, raw = TRUE)3 0.00007018 0.00001859 3.775 0.00003374 0.0001066 fixed
male 0.0001426 0.02365 0.006029 -0.04621 0.0465 fixed
sibling_count3 -0.02208 0.04182 -0.5281 -0.104 0.05988 fixed
sibling_count4 0.0252 0.04553 0.5535 -0.06404 0.1144 fixed
sibling_count5 -0.04036 0.05006 -0.8061 -0.1385 0.05777 fixed
sibling_count5+ -0.08923 0.04936 -1.808 -0.186 0.007504 fixed
birth_order_nonlinear2 -0.007419 0.03111 -0.2385 -0.06839 0.05355 fixed
birth_order_nonlinear3 -0.06747 0.03749 -1.8 -0.1409 0.006012 fixed
birth_order_nonlinear4 -0.08405 0.04525 -1.857 -0.1728 0.004641 fixed
birth_order_nonlinear5 0.01327 0.05537 0.2397 -0.09524 0.1218 fixed
birth_order_nonlinear5+ -0.1017 0.05263 -1.933 -0.2049 0.001402 fixed
sd_(Intercept).mother_pidlink 0.3854 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8342 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.753 0.4063 -4.315 -2.55 -0.9569 fixed
poly(age, 3, raw = TRUE)1 0.2086 0.04603 4.531 0.1183 0.2988 fixed
poly(age, 3, raw = TRUE)2 -0.006848 0.001643 -4.168 -0.01007 -0.003628 fixed
poly(age, 3, raw = TRUE)3 0.00006923 0.00001861 3.72 0.00003276 0.0001057 fixed
male -0.0008112 0.02365 -0.03431 -0.04715 0.04553 fixed
count_birth_order2/2 0.1046 0.0606 1.727 -0.01412 0.2234 fixed
count_birth_order1/3 0.04638 0.05345 0.8677 -0.05839 0.1511 fixed
count_birth_order2/3 -0.03838 0.05776 -0.6645 -0.1516 0.07483 fixed
count_birth_order3/3 -0.04669 0.06486 -0.7199 -0.1738 0.08044 fixed
count_birth_order1/4 0.08308 0.06328 1.313 -0.04095 0.2071 fixed
count_birth_order2/4 0.1233 0.06453 1.91 -0.003195 0.2498 fixed
count_birth_order3/4 -0.09541 0.07025 -1.358 -0.2331 0.04228 fixed
count_birth_order4/4 -0.0516 0.07268 -0.71 -0.194 0.09084 fixed
count_birth_order1/5 -0.01385 0.07452 -0.1859 -0.1599 0.1322 fixed
count_birth_order2/5 -0.068 0.08025 -0.8473 -0.2253 0.08929 fixed
count_birth_order3/5 -0.08213 0.078 -1.053 -0.235 0.07074 fixed
count_birth_order4/5 -0.01231 0.08102 -0.1519 -0.1711 0.1465 fixed
count_birth_order5/5 0.02339 0.08067 0.29 -0.1347 0.1815 fixed
count_birth_order1/5+ -0.04074 0.07093 -0.5743 -0.1798 0.09829 fixed
count_birth_order2/5+ -0.1348 0.0748 -1.802 -0.2814 0.01179 fixed
count_birth_order3/5+ -0.02254 0.07276 -0.3097 -0.1651 0.1201 fixed
count_birth_order4/5+ -0.1605 0.07 -2.292 -0.2977 -0.02327 fixed
count_birth_order5/5+ -0.04646 0.07191 -0.6462 -0.1874 0.09447 fixed
count_birth_order5+/5+ -0.1536 0.05342 -2.876 -0.2583 -0.04895 fixed
sd_(Intercept).mother_pidlink 0.3886 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8324 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 15665 15738 -7821 15643 NA NA NA
12 15664 15744 -7820 15640 2.876 1 0.08993
16 15665 15772 -7817 15633 6.217 4 0.1835
26 15669 15843 -7809 15617 16.19 10 0.09423

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.758 0.4107 -4.281 -2.563 -0.9531 fixed
poly(age, 3, raw = TRUE)1 0.2094 0.0466 4.494 0.1181 0.3007 fixed
poly(age, 3, raw = TRUE)2 -0.006886 0.001663 -4.142 -0.01014 -0.003627 fixed
poly(age, 3, raw = TRUE)3 0.00007041 0.00001883 3.738 0.00003349 0.0001073 fixed
male -0.001767 0.02395 -0.07378 -0.04872 0.04518 fixed
sibling_count3 0.02318 0.03734 0.6207 -0.05001 0.09636 fixed
sibling_count4 -0.01709 0.04066 -0.4203 -0.09679 0.0626 fixed
sibling_count5 -0.03092 0.04791 -0.6453 -0.1248 0.06299 fixed
sibling_count5+ -0.1816 0.04168 -4.358 -0.2633 -0.09994 fixed
sd_(Intercept).mother_pidlink 0.3835 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8335 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.751 0.4107 -4.263 -2.556 -0.9461 fixed
birth_order -0.009102 0.00815 -1.117 -0.02508 0.006872 fixed
poly(age, 3, raw = TRUE)1 0.2099 0.0466 4.505 0.1186 0.3012 fixed
poly(age, 3, raw = TRUE)2 -0.006889 0.001663 -4.144 -0.01015 -0.003631 fixed
poly(age, 3, raw = TRUE)3 0.00007008 0.00001884 3.72 0.00003316 0.000107 fixed
male -0.001512 0.02396 -0.06313 -0.04846 0.04544 fixed
sibling_count3 0.02767 0.03755 0.737 -0.04592 0.1013 fixed
sibling_count4 -0.006581 0.04173 -0.1577 -0.08836 0.0752 fixed
sibling_count5 -0.01437 0.05014 -0.2867 -0.1126 0.0839 fixed
sibling_count5+ -0.1474 0.05172 -2.85 -0.2488 -0.04602 fixed
sd_(Intercept).mother_pidlink 0.3829 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8337 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.75 0.4113 -4.255 -2.556 -0.9438 fixed
poly(age, 3, raw = TRUE)1 0.2089 0.04663 4.48 0.1175 0.3003 fixed
poly(age, 3, raw = TRUE)2 -0.006853 0.001664 -4.119 -0.01011 -0.003592 fixed
poly(age, 3, raw = TRUE)3 0.00006965 0.00001885 3.695 0.00003271 0.0001066 fixed
male -0.001605 0.02396 -0.06699 -0.04856 0.04535 fixed
sibling_count3 0.03932 0.03833 1.026 -0.03581 0.1144 fixed
sibling_count4 0.006506 0.04337 0.15 -0.0785 0.09151 fixed
sibling_count5 -0.001042 0.05236 -0.01989 -0.1037 0.1016 fixed
sibling_count5+ -0.1381 0.05322 -2.595 -0.2424 -0.03379 fixed
birth_order_nonlinear2 -0.01308 0.03042 -0.4302 -0.0727 0.04653 fixed
birth_order_nonlinear3 -0.06863 0.03759 -1.826 -0.1423 0.005042 fixed
birth_order_nonlinear4 -0.03461 0.04784 -0.7235 -0.1284 0.05915 fixed
birth_order_nonlinear5 -0.0511 0.06064 -0.8426 -0.17 0.06776 fixed
birth_order_nonlinear5+ -0.06211 0.06069 -1.023 -0.181 0.05683 fixed
sd_(Intercept).mother_pidlink 0.384 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8334 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.72 0.4121 -4.173 -2.527 -0.9121 fixed
poly(age, 3, raw = TRUE)1 0.2027 0.04671 4.339 0.1111 0.2942 fixed
poly(age, 3, raw = TRUE)2 -0.006614 0.001667 -3.968 -0.009881 -0.003347 fixed
poly(age, 3, raw = TRUE)3 0.00006677 0.00001889 3.535 0.00002975 0.0001038 fixed
male -0.004393 0.02396 -0.1833 -0.05136 0.04257 fixed
count_birth_order2/2 0.05125 0.0538 0.9527 -0.05419 0.1567 fixed
count_birth_order1/3 0.06193 0.04865 1.273 -0.03342 0.1573 fixed
count_birth_order2/3 0.0383 0.05364 0.7141 -0.06683 0.1434 fixed
count_birth_order3/3 0.00373 0.05886 0.06337 -0.1116 0.1191 fixed
count_birth_order1/4 0.008194 0.06064 0.1351 -0.1107 0.127 fixed
count_birth_order2/4 0.05461 0.06234 0.8759 -0.06759 0.1768 fixed
count_birth_order3/4 -0.07707 0.06483 -1.189 -0.2041 0.05 fixed
count_birth_order4/4 0.01361 0.06861 0.1983 -0.1209 0.1481 fixed
count_birth_order1/5 0.06726 0.08109 0.8294 -0.09168 0.2262 fixed
count_birth_order2/5 -0.03652 0.09006 -0.4055 -0.213 0.14 fixed
count_birth_order3/5 -0.1304 0.0861 -1.515 -0.2992 0.03832 fixed
count_birth_order4/5 0.03388 0.0827 0.4096 -0.1282 0.196 fixed
count_birth_order5/5 -0.01935 0.08732 -0.2216 -0.1905 0.1518 fixed
count_birth_order1/5+ -0.01061 0.08265 -0.1284 -0.1726 0.1514 fixed
count_birth_order2/5+ -0.2631 0.08334 -3.158 -0.4265 -0.09981 fixed
count_birth_order3/5+ -0.07723 0.0815 -0.9476 -0.237 0.08251 fixed
count_birth_order4/5+ -0.2221 0.07927 -2.802 -0.3775 -0.06677 fixed
count_birth_order5/5+ -0.174 0.07346 -2.368 -0.318 -0.03 fixed
count_birth_order5+/5+ -0.1797 0.05555 -3.235 -0.2886 -0.07082 fixed
sd_(Intercept).mother_pidlink 0.3852 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8326 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 15200 15273 -7589 15178 NA NA NA
12 15200 15280 -7588 15176 1.251 1 0.2634
16 15206 15312 -7587 15174 2.534 4 0.6385
26 15211 15384 -7580 15159 14.26 10 0.1615

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Immediate word recall

birthorder <- birthorder %>% mutate(outcome = words_immediate)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.3533 0.1522 2.322 0.05506 0.6515 fixed
poly(age, 3, raw = TRUE)1 0.006171 0.01459 0.4228 -0.02243 0.03478 fixed
poly(age, 3, raw = TRUE)2 -0.0004583 0.0004298 -1.066 -0.001301 0.0003842 fixed
poly(age, 3, raw = TRUE)3 0.0000006377 0.000003962 0.161 -0.000007128 0.000008403 fixed
male -0.1026 0.01588 -6.461 -0.1337 -0.07148 fixed
sibling_count3 0.02844 0.03284 0.8661 -0.03592 0.0928 fixed
sibling_count4 0.0288 0.03393 0.8488 -0.0377 0.0953 fixed
sibling_count5 -0.005082 0.03538 -0.1436 -0.07443 0.06426 fixed
sibling_count5+ -0.1109 0.02768 -4.006 -0.1651 -0.05663 fixed
sd_(Intercept).mother_pidlink 0.3956 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8692 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.3527 0.1522 2.318 0.05448 0.651 fixed
birth_order -0.002253 0.003393 -0.6641 -0.008902 0.004396 fixed
poly(age, 3, raw = TRUE)1 0.006857 0.01463 0.4687 -0.02182 0.03553 fixed
poly(age, 3, raw = TRUE)2 -0.0004838 0.0004315 -1.121 -0.00133 0.000362 fixed
poly(age, 3, raw = TRUE)3 0.0000008756 0.000003978 0.2201 -0.000006921 0.000008673 fixed
male -0.1025 0.01588 -6.457 -0.1337 -0.07142 fixed
sibling_count3 0.02891 0.03285 0.8802 -0.03547 0.09329 fixed
sibling_count4 0.03031 0.034 0.8913 -0.03634 0.09695 fixed
sibling_count5 -0.002406 0.03561 -0.06757 -0.0722 0.06738 fixed
sibling_count5+ -0.1024 0.03047 -3.361 -0.1621 -0.04269 fixed
sd_(Intercept).mother_pidlink 0.3954 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8693 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.3551 0.1526 2.327 0.05601 0.6541 fixed
poly(age, 3, raw = TRUE)1 0.007002 0.01463 0.4784 -0.02168 0.03569 fixed
poly(age, 3, raw = TRUE)2 -0.0004887 0.0004316 -1.132 -0.001335 0.0003572 fixed
poly(age, 3, raw = TRUE)3 0.0000009008 0.00000398 0.2264 -0.000006899 0.000008701 fixed
male -0.1025 0.01588 -6.451 -0.1336 -0.07133 fixed
sibling_count3 0.02886 0.03328 0.8672 -0.03637 0.09408 fixed
sibling_count4 0.02953 0.0349 0.8461 -0.03887 0.09792 fixed
sibling_count5 -0.00005737 0.03689 -0.001555 -0.07236 0.07224 fixed
sibling_count5+ -0.09666 0.03193 -3.027 -0.1592 -0.03408 fixed
birth_order_nonlinear2 -0.01544 0.02294 -0.6729 -0.0604 0.02953 fixed
birth_order_nonlinear3 -0.008737 0.02705 -0.323 -0.06175 0.04428 fixed
birth_order_nonlinear4 -0.005941 0.03082 -0.1928 -0.06635 0.05446 fixed
birth_order_nonlinear5 -0.03387 0.03506 -0.9661 -0.1026 0.03485 fixed
birth_order_nonlinear5+ -0.03147 0.02934 -1.073 -0.08897 0.02603 fixed
sd_(Intercept).mother_pidlink 0.3954 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8694 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.3423 0.1532 2.234 0.04197 0.6426 fixed
poly(age, 3, raw = TRUE)1 0.007367 0.01464 0.5033 -0.02132 0.03606 fixed
poly(age, 3, raw = TRUE)2 -0.0005014 0.0004317 -1.161 -0.001348 0.0003448 fixed
poly(age, 3, raw = TRUE)3 0.000001029 0.000003981 0.2583 -0.000006775 0.000008832 fixed
male -0.1027 0.01589 -6.463 -0.1338 -0.07155 fixed
count_birth_order2/2 0.01152 0.04462 0.2582 -0.07594 0.09898 fixed
count_birth_order1/3 0.03433 0.04294 0.7994 -0.04984 0.1185 fixed
count_birth_order2/3 0.03836 0.04795 0.8002 -0.05561 0.1323 fixed
count_birth_order3/3 0.01806 0.05364 0.3367 -0.08707 0.1232 fixed
count_birth_order1/4 0.009683 0.04898 0.1977 -0.08632 0.1057 fixed
count_birth_order2/4 0.06327 0.05141 1.231 -0.03749 0.164 fixed
count_birth_order3/4 0.0178 0.0558 0.3189 -0.09158 0.1272 fixed
count_birth_order4/4 0.04206 0.05908 0.7118 -0.07374 0.1579 fixed
count_birth_order1/5 0.06302 0.05549 1.136 -0.04573 0.1718 fixed
count_birth_order2/5 -0.05377 0.05819 -0.9241 -0.1678 0.06027 fixed
count_birth_order3/5 -0.02803 0.05982 -0.4686 -0.1453 0.08922 fixed
count_birth_order4/5 -0.007194 0.06332 -0.1136 -0.1313 0.1169 fixed
count_birth_order5/5 0.007012 0.06469 0.1084 -0.1198 0.1338 fixed
count_birth_order1/5+ -0.07213 0.0447 -1.614 -0.1597 0.01547 fixed
count_birth_order2/5+ -0.135 0.04606 -2.931 -0.2253 -0.04471 fixed
count_birth_order3/5+ -0.06838 0.04515 -1.514 -0.1569 0.02012 fixed
count_birth_order4/5+ -0.09254 0.04425 -2.091 -0.1793 -0.005813 fixed
count_birth_order5/5+ -0.1312 0.04458 -2.943 -0.2185 -0.04381 fixed
count_birth_order5+/5+ -0.1183 0.03526 -3.355 -0.1874 -0.04921 fixed
sd_(Intercept).mother_pidlink 0.3952 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8695 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 38025 38108 -19002 38003 NA NA NA
12 38027 38117 -19001 38003 0.4421 1 0.5061
16 38034 38154 -19001 38002 1.325 4 0.8572
26 38046 38242 -18997 37994 7.138 10 0.7124

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.2703 0.4011 0.6739 -0.5158 1.056 fixed
poly(age, 3, raw = TRUE)1 0.0125 0.04552 0.2745 -0.07672 0.1017 fixed
poly(age, 3, raw = TRUE)2 -0.0003482 0.001625 -0.2144 -0.003532 0.002836 fixed
poly(age, 3, raw = TRUE)3 0.000000429 0.0000184 0.02331 -0.00003564 0.00003649 fixed
male -0.1315 0.02342 -5.616 -0.1774 -0.08563 fixed
sibling_count3 0.03636 0.03708 0.9806 -0.03632 0.109 fixed
sibling_count4 -0.08697 0.03999 -2.175 -0.1653 -0.008589 fixed
sibling_count5 -0.1077 0.04571 -2.357 -0.1973 -0.01814 fixed
sibling_count5+ -0.1732 0.04014 -4.315 -0.2519 -0.09454 fixed
sd_(Intercept).mother_pidlink 0.3506 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8393 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.2706 0.4012 0.6746 -0.5157 1.057 fixed
birth_order -0.0004093 0.007764 -0.05271 -0.01563 0.01481 fixed
poly(age, 3, raw = TRUE)1 0.01252 0.04552 0.2749 -0.07671 0.1017 fixed
poly(age, 3, raw = TRUE)2 -0.0003483 0.001625 -0.2144 -0.003533 0.002836 fixed
poly(age, 3, raw = TRUE)3 0.0000004136 0.0000184 0.02247 -0.00003566 0.00003649 fixed
male -0.1315 0.02342 -5.614 -0.1774 -0.0856 fixed
sibling_count3 0.03656 0.03728 0.9808 -0.0365 0.1096 fixed
sibling_count4 -0.08649 0.04099 -2.11 -0.1668 -0.006159 fixed
sibling_count5 -0.107 0.04802 -2.227 -0.2011 -0.01284 fixed
sibling_count5+ -0.1717 0.04976 -3.45 -0.2692 -0.07415 fixed
sd_(Intercept).mother_pidlink 0.3506 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8393 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.3012 0.4018 0.7496 -0.4863 1.089 fixed
poly(age, 3, raw = TRUE)1 0.01015 0.04555 0.2228 -0.07912 0.09942 fixed
poly(age, 3, raw = TRUE)2 -0.0002567 0.001626 -0.1579 -0.003443 0.002929 fixed
poly(age, 3, raw = TRUE)3 -0.0000007896 0.00001842 -0.04287 -0.00003689 0.00003531 fixed
male -0.1317 0.02342 -5.625 -0.1777 -0.08584 fixed
sibling_count3 0.03149 0.03803 0.8279 -0.04305 0.106 fixed
sibling_count4 -0.08309 0.04257 -1.952 -0.1665 0.0003424 fixed
sibling_count5 -0.08576 0.05042 -1.701 -0.1846 0.01307 fixed
sibling_count5+ -0.1567 0.05121 -3.06 -0.2571 -0.05632 fixed
birth_order_nonlinear2 -0.03205 0.03018 -1.062 -0.0912 0.02709 fixed
birth_order_nonlinear3 0.02066 0.03726 0.5544 -0.05237 0.09368 fixed
birth_order_nonlinear4 -0.04889 0.04612 -1.06 -0.1393 0.0415 fixed
birth_order_nonlinear5 -0.09808 0.05756 -1.704 -0.2109 0.01473 fixed
birth_order_nonlinear5+ -0.00781 0.05794 -0.1348 -0.1214 0.1057 fixed
sd_(Intercept).mother_pidlink 0.3513 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8389 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.326 0.4027 0.8094 -0.4633 1.115 fixed
poly(age, 3, raw = TRUE)1 0.007183 0.04565 0.1573 -0.08229 0.09665 fixed
poly(age, 3, raw = TRUE)2 -0.0001525 0.00163 -0.09357 -0.003346 0.003041 fixed
poly(age, 3, raw = TRUE)3 -0.00000193 0.00001846 -0.1045 -0.00003812 0.00003426 fixed
male -0.1314 0.02344 -5.606 -0.1774 -0.08548 fixed
count_birth_order2/2 -0.02869 0.05464 -0.5251 -0.1358 0.0784 fixed
count_birth_order1/3 0.03888 0.04861 0.7998 -0.0564 0.1342 fixed
count_birth_order2/3 0.0252 0.05306 0.475 -0.0788 0.1292 fixed
count_birth_order3/3 0.006071 0.05925 0.1025 -0.1101 0.1222 fixed
count_birth_order1/4 -0.09194 0.05943 -1.547 -0.2084 0.02455 fixed
count_birth_order2/4 -0.1157 0.0614 -1.883 -0.236 0.004696 fixed
count_birth_order3/4 -0.0483 0.06482 -0.7451 -0.1754 0.07875 fixed
count_birth_order4/4 -0.1287 0.06748 -1.907 -0.261 0.003563 fixed
count_birth_order1/5 -0.1045 0.08064 -1.296 -0.2625 0.05355 fixed
count_birth_order2/5 -0.07683 0.08625 -0.8907 -0.2459 0.09223 fixed
count_birth_order3/5 -0.06136 0.08085 -0.7589 -0.2198 0.09711 fixed
count_birth_order4/5 -0.1804 0.0779 -2.316 -0.3331 -0.02774 fixed
count_birth_order5/5 -0.1471 0.08104 -1.816 -0.306 0.01169 fixed
count_birth_order1/5+ -0.1353 0.07997 -1.692 -0.2921 0.02142 fixed
count_birth_order2/5+ -0.2916 0.07915 -3.684 -0.4467 -0.1365 fixed
count_birth_order3/5+ -0.06251 0.07924 -0.7889 -0.2178 0.09279 fixed
count_birth_order4/5+ -0.1658 0.07465 -2.221 -0.3121 -0.01951 fixed
count_birth_order5/5+ -0.2781 0.07055 -3.942 -0.4164 -0.1399 fixed
count_birth_order5+/5+ -0.1635 0.05347 -3.058 -0.2683 -0.05874 fixed
sd_(Intercept).mother_pidlink 0.3506 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8395 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 15660 15733 -7819 15638 NA NA NA
12 15662 15742 -7819 15638 0.002809 1 0.9577
16 15664 15771 -7816 15632 6.129 4 0.1897
26 15677 15851 -7813 15625 6.632 10 0.7597

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.3023 0.4002 0.7553 -0.4821 1.087 fixed
poly(age, 3, raw = TRUE)1 0.00985 0.04544 0.2168 -0.0792 0.0989 fixed
poly(age, 3, raw = TRUE)2 -0.0002896 0.001622 -0.1786 -0.003469 0.00289 fixed
poly(age, 3, raw = TRUE)3 -0.0000002249 0.00001838 -0.01224 -0.00003624 0.00003579 fixed
male -0.1318 0.02335 -5.645 -0.1775 -0.08603 fixed
sibling_count3 0.03106 0.04014 0.7739 -0.04761 0.1097 fixed
sibling_count4 -0.06748 0.04232 -1.595 -0.1504 0.01547 fixed
sibling_count5 -0.05013 0.04535 -1.105 -0.139 0.03875 fixed
sibling_count5+ -0.1294 0.03973 -3.257 -0.2073 -0.05154 fixed
sd_(Intercept).mother_pidlink 0.3519 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8397 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.3119 0.4003 0.7792 -0.4726 1.096 fixed
birth_order -0.008494 0.00679 -1.251 -0.0218 0.004814 fixed
poly(age, 3, raw = TRUE)1 0.009959 0.04544 0.2192 -0.07909 0.09901 fixed
poly(age, 3, raw = TRUE)2 -0.0002798 0.001622 -0.1725 -0.003459 0.002899 fixed
poly(age, 3, raw = TRUE)3 -0.0000006759 0.00001838 -0.03678 -0.0000367 0.00003534 fixed
male -0.1315 0.02335 -5.631 -0.1772 -0.08571 fixed
sibling_count3 0.03519 0.04027 0.8737 -0.04375 0.1141 fixed
sibling_count4 -0.05807 0.04298 -1.351 -0.1423 0.02617 fixed
sibling_count5 -0.03529 0.04687 -0.7529 -0.1272 0.05658 fixed
sibling_count5+ -0.09817 0.04692 -2.092 -0.1901 -0.006204 fixed
sd_(Intercept).mother_pidlink 0.3517 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8398 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.3142 0.4008 0.7839 -0.4713 1.1 fixed
poly(age, 3, raw = TRUE)1 0.009226 0.04545 0.203 -0.07986 0.09831 fixed
poly(age, 3, raw = TRUE)2 -0.0002426 0.001623 -0.1495 -0.003423 0.002938 fixed
poly(age, 3, raw = TRUE)3 -0.000001365 0.00001839 -0.07425 -0.00003741 0.00003468 fixed
male -0.1316 0.02335 -5.636 -0.1774 -0.08583 fixed
sibling_count3 0.03017 0.04099 0.7361 -0.05017 0.1105 fixed
sibling_count4 -0.05283 0.04447 -1.188 -0.14 0.03434 fixed
sibling_count5 -0.02045 0.04908 -0.4166 -0.1167 0.07576 fixed
sibling_count5+ -0.07574 0.04836 -1.566 -0.1705 0.01904 fixed
birth_order_nonlinear2 -0.02291 0.03076 -0.7446 -0.0832 0.03739 fixed
birth_order_nonlinear3 0.00622 0.03718 0.1673 -0.06665 0.07909 fixed
birth_order_nonlinear4 -0.07617 0.04479 -1.701 -0.164 0.01162 fixed
birth_order_nonlinear5 -0.09325 0.05494 -1.697 -0.2009 0.01442 fixed
birth_order_nonlinear5+ -0.09131 0.05184 -1.761 -0.1929 0.01029 fixed
sd_(Intercept).mother_pidlink 0.353 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8392 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.3186 0.4016 0.7935 -0.4684 1.106 fixed
poly(age, 3, raw = TRUE)1 0.00678 0.04552 0.1489 -0.08245 0.096 fixed
poly(age, 3, raw = TRUE)2 -0.0001478 0.001626 -0.09093 -0.003334 0.003038 fixed
poly(age, 3, raw = TRUE)3 -0.000002497 0.00001842 -0.1355 -0.00003861 0.00003361 fixed
male -0.1314 0.02337 -5.623 -0.1772 -0.08558 fixed
count_birth_order2/2 0.02163 0.05991 0.361 -0.09579 0.139 fixed
count_birth_order1/3 0.05145 0.05267 0.9768 -0.05178 0.1547 fixed
count_birth_order2/3 0.03833 0.05711 0.6712 -0.0736 0.1503 fixed
count_birth_order3/3 0.01506 0.06423 0.2345 -0.1108 0.1409 fixed
count_birth_order1/4 -0.06686 0.06228 -1.073 -0.1889 0.05522 fixed
count_birth_order2/4 -0.05847 0.06323 -0.9247 -0.1824 0.06546 fixed
count_birth_order3/4 -0.03496 0.06945 -0.5034 -0.1711 0.1012 fixed
count_birth_order4/4 -0.06921 0.07172 -0.965 -0.2098 0.07136 fixed
count_birth_order1/5 0.02871 0.07388 0.3886 -0.1161 0.1735 fixed
count_birth_order2/5 0.0003236 0.07927 0.004082 -0.155 0.1557 fixed
count_birth_order3/5 -0.009014 0.07708 -0.1169 -0.1601 0.1421 fixed
count_birth_order4/5 -0.09693 0.07987 -1.214 -0.2535 0.05961 fixed
count_birth_order5/5 -0.1483 0.08009 -1.851 -0.3052 0.008712 fixed
count_birth_order1/5+ -0.02559 0.07021 -0.3644 -0.1632 0.112 fixed
count_birth_order2/5+ -0.1887 0.07328 -2.575 -0.3324 -0.04509 fixed
count_birth_order3/5+ 0.00315 0.07175 0.04391 -0.1375 0.1438 fixed
count_birth_order4/5+ -0.1667 0.06916 -2.41 -0.3023 -0.03115 fixed
count_birth_order5/5+ -0.1191 0.07104 -1.676 -0.2583 0.02015 fixed
count_birth_order5+/5+ -0.1524 0.0524 -2.909 -0.2551 -0.04972 fixed
sd_(Intercept).mother_pidlink 0.3507 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8402 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 15806 15880 -7892 15784 NA NA NA
12 15807 15887 -7891 15783 1.568 1 0.2105
16 15810 15917 -7889 15778 5.324 4 0.2556
26 15821 15996 -7885 15769 8.136 10 0.6156

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.2817 0.4058 0.6942 -0.5136 1.077 fixed
poly(age, 3, raw = TRUE)1 0.0102 0.04607 0.2214 -0.08009 0.1005 fixed
poly(age, 3, raw = TRUE)2 -0.0002244 0.001645 -0.1364 -0.003448 0.002999 fixed
poly(age, 3, raw = TRUE)3 -0.000001458 0.00001864 -0.0782 -0.00003799 0.00003507 fixed
male -0.127 0.02366 -5.368 -0.1734 -0.08062 fixed
sibling_count3 0.01622 0.03661 0.443 -0.05554 0.08798 fixed
sibling_count4 -0.08336 0.03974 -2.098 -0.1612 -0.005483 fixed
sibling_count5 -0.1066 0.04688 -2.274 -0.1985 -0.01472 fixed
sibling_count5+ -0.1814 0.04062 -4.467 -0.261 -0.1018 fixed
sd_(Intercept).mother_pidlink 0.3519 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8386 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.2826 0.4058 0.6965 -0.5128 1.078 fixed
birth_order -0.001434 0.007998 -0.1793 -0.01711 0.01424 fixed
poly(age, 3, raw = TRUE)1 0.01029 0.04608 0.2234 -0.08001 0.1006 fixed
poly(age, 3, raw = TRUE)2 -0.0002252 0.001645 -0.1369 -0.003449 0.002999 fixed
poly(age, 3, raw = TRUE)3 -0.000001507 0.00001864 -0.08082 -0.00003805 0.00003503 fixed
male -0.127 0.02366 -5.365 -0.1733 -0.08058 fixed
sibling_count3 0.01692 0.03683 0.4595 -0.05525 0.0891 fixed
sibling_count4 -0.08173 0.04077 -2.005 -0.1616 -0.001818 fixed
sibling_count5 -0.104 0.04907 -2.12 -0.2002 -0.007835 fixed
sibling_count5+ -0.1761 0.05051 -3.486 -0.2751 -0.07706 fixed
sd_(Intercept).mother_pidlink 0.3519 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8387 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.3087 0.4063 0.7598 -0.4876 1.105 fixed
poly(age, 3, raw = TRUE)1 0.008334 0.04608 0.1809 -0.08199 0.09866 fixed
poly(age, 3, raw = TRUE)2 -0.0001465 0.001645 -0.08904 -0.003371 0.003078 fixed
poly(age, 3, raw = TRUE)3 -0.000002585 0.00001865 -0.1386 -0.00003914 0.00003397 fixed
male -0.1268 0.02365 -5.361 -0.1732 -0.08045 fixed
sibling_count3 0.007929 0.03761 0.2108 -0.06578 0.08164 fixed
sibling_count4 -0.08449 0.04241 -1.992 -0.1676 -0.001375 fixed
sibling_count5 -0.08849 0.05131 -1.724 -0.1891 0.01209 fixed
sibling_count5+ -0.1624 0.05206 -3.12 -0.2644 -0.06038 fixed
birth_order_nonlinear2 -0.03418 0.03009 -1.136 -0.09315 0.0248 fixed
birth_order_nonlinear3 0.03246 0.03725 0.8714 -0.04054 0.1055 fixed
birth_order_nonlinear4 -0.04026 0.04743 -0.8487 -0.1332 0.05271 fixed
birth_order_nonlinear5 -0.1089 0.06002 -1.814 -0.2265 0.008782 fixed
birth_order_nonlinear5+ -0.01919 0.05976 -0.3212 -0.1363 0.09792 fixed
sd_(Intercept).mother_pidlink 0.3542 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8375 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.3267 0.4074 0.8019 -0.4718 1.125 fixed
poly(age, 3, raw = TRUE)1 0.006268 0.04621 0.1357 -0.08429 0.09683 fixed
poly(age, 3, raw = TRUE)2 -0.00007614 0.00165 -0.04615 -0.00331 0.003158 fixed
poly(age, 3, raw = TRUE)3 -0.000003334 0.0000187 -0.1783 -0.00003999 0.00003333 fixed
male -0.1272 0.02368 -5.373 -0.1736 -0.08082 fixed
count_birth_order2/2 -0.03157 0.05319 -0.5935 -0.1358 0.07268 fixed
count_birth_order1/3 0.00815 0.04803 0.1697 -0.08598 0.1023 fixed
count_birth_order2/3 -0.003243 0.05303 -0.06115 -0.1072 0.1007 fixed
count_birth_order3/3 0.01332 0.0582 0.2289 -0.1007 0.1274 fixed
count_birth_order1/4 -0.07328 0.05972 -1.227 -0.1903 0.04377 fixed
count_birth_order2/4 -0.1149 0.0613 -1.874 -0.235 0.005281 fixed
count_birth_order3/4 -0.05864 0.06415 -0.9141 -0.1844 0.06709 fixed
count_birth_order4/4 -0.1343 0.06788 -1.978 -0.2673 -0.001241 fixed
count_birth_order1/5 -0.112 0.08048 -1.391 -0.2697 0.04577 fixed
count_birth_order2/5 -0.09691 0.08883 -1.091 -0.271 0.07719 fixed
count_birth_order3/5 -0.05437 0.08475 -0.6416 -0.2205 0.1117 fixed
count_birth_order4/5 -0.1419 0.08168 -1.738 -0.302 0.01816 fixed
count_birth_order5/5 -0.174 0.08675 -2.006 -0.344 -0.003991 fixed
count_birth_order1/5+ -0.1511 0.08185 -1.846 -0.3115 0.009357 fixed
count_birth_order2/5+ -0.2914 0.08114 -3.591 -0.4504 -0.1324 fixed
count_birth_order3/5+ -0.05562 0.08034 -0.6923 -0.2131 0.1019 fixed
count_birth_order4/5+ -0.1748 0.07855 -2.226 -0.3288 -0.02089 fixed
count_birth_order5/5+ -0.2837 0.07222 -3.929 -0.4253 -0.1422 fixed
count_birth_order5+/5+ -0.1808 0.05448 -3.319 -0.2876 -0.07403 fixed
sd_(Intercept).mother_pidlink 0.3538 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8382 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 15343 15417 -7661 15321 NA NA NA
12 15345 15425 -7661 15321 0.03222 1 0.8575
16 15346 15453 -7657 15314 7.259 4 0.1228
26 15362 15535 -7655 15310 4.273 10 0.9342

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Delayed word recall

birthorder <- birthorder %>% mutate(outcome = words_delayed)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.309 0.1522 2.03 0.01067 0.6073 fixed
poly(age, 3, raw = TRUE)1 0.01178 0.0146 0.8074 -0.01682 0.04039 fixed
poly(age, 3, raw = TRUE)2 -0.0006651 0.0004297 -1.548 -0.001507 0.0001771 fixed
poly(age, 3, raw = TRUE)3 0.000001695 0.00000396 0.4279 -0.000006066 0.000009455 fixed
male -0.07984 0.01592 -5.015 -0.111 -0.04864 fixed
sibling_count3 0.05182 0.03269 1.585 -0.01226 0.1159 fixed
sibling_count4 0.03874 0.03375 1.148 -0.02741 0.1049 fixed
sibling_count5 0.007162 0.03517 0.2037 -0.06176 0.07609 fixed
sibling_count5+ -0.06697 0.02754 -2.432 -0.1209 -0.01299 fixed
sd_(Intercept).mother_pidlink 0.3767 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8767 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.3084 0.1522 2.026 0.01012 0.6068 fixed
birth_order -0.001883 0.003389 -0.5557 -0.008525 0.004759 fixed
poly(age, 3, raw = TRUE)1 0.01235 0.01463 0.8444 -0.01632 0.04103 fixed
poly(age, 3, raw = TRUE)2 -0.0006861 0.0004314 -1.591 -0.001532 0.0001594 fixed
poly(age, 3, raw = TRUE)3 0.00000189 0.000003975 0.4754 -0.000005902 0.000009681 fixed
male -0.07979 0.01592 -5.012 -0.111 -0.04858 fixed
sibling_count3 0.05222 0.0327 1.597 -0.01188 0.1163 fixed
sibling_count4 0.04001 0.03383 1.183 -0.02629 0.1063 fixed
sibling_count5 0.00941 0.0354 0.2658 -0.05997 0.07879 fixed
sibling_count5+ -0.05988 0.03035 -1.973 -0.1194 -0.0003944 fixed
sd_(Intercept).mother_pidlink 0.3766 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8767 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.3002 0.1526 1.967 0.001069 0.5993 fixed
poly(age, 3, raw = TRUE)1 0.01267 0.01463 0.8656 -0.01602 0.04135 fixed
poly(age, 3, raw = TRUE)2 -0.0007038 0.0004314 -1.631 -0.001549 0.0001418 fixed
poly(age, 3, raw = TRUE)3 0.000002105 0.000003976 0.5294 -0.000005689 0.000009899 fixed
male -0.07975 0.01592 -5.009 -0.111 -0.04854 fixed
sibling_count3 0.054 0.03314 1.63 -0.01095 0.119 fixed
sibling_count4 0.03784 0.03474 1.089 -0.03024 0.1059 fixed
sibling_count5 0.00649 0.03669 0.1769 -0.06543 0.07841 fixed
sibling_count5+ -0.05398 0.03183 -1.696 -0.1164 0.008402 fixed
birth_order_nonlinear2 0.01565 0.02303 0.6794 -0.02949 0.06078 fixed
birth_order_nonlinear3 -0.008068 0.02716 -0.2971 -0.0613 0.04516 fixed
birth_order_nonlinear4 0.01986 0.03094 0.6418 -0.04079 0.08051 fixed
birth_order_nonlinear5 0.005391 0.0352 0.1531 -0.06361 0.07439 fixed
birth_order_nonlinear5+ -0.02652 0.02938 -0.9024 -0.0841 0.03107 fixed
sd_(Intercept).mother_pidlink 0.3763 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8769 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.2999 0.1533 1.957 -0.0005013 0.6003 fixed
poly(age, 3, raw = TRUE)1 0.0129 0.01464 0.8813 -0.01579 0.04159 fixed
poly(age, 3, raw = TRUE)2 -0.0007085 0.0004316 -1.642 -0.001554 0.0001374 fixed
poly(age, 3, raw = TRUE)3 0.000002131 0.000003979 0.5357 -0.000005667 0.000009929 fixed
male -0.07974 0.01593 -5.007 -0.111 -0.04853 fixed
count_birth_order2/2 0.008045 0.04478 0.1796 -0.07973 0.09582 fixed
count_birth_order1/3 0.06475 0.04293 1.508 -0.01939 0.1489 fixed
count_birth_order2/3 0.05148 0.04795 1.074 -0.0425 0.1455 fixed
count_birth_order3/3 0.03626 0.05366 0.6758 -0.06891 0.1414 fixed
count_birth_order1/4 0.006214 0.04899 0.1268 -0.0898 0.1022 fixed
count_birth_order2/4 0.1041 0.05142 2.024 0.003272 0.2048 fixed
count_birth_order3/4 0.02208 0.05583 0.3954 -0.08735 0.1315 fixed
count_birth_order4/4 0.03046 0.05912 0.5152 -0.08542 0.1463 fixed
count_birth_order1/5 -0.04227 0.05551 -0.7614 -0.1511 0.06654 fixed
count_birth_order2/5 0.02579 0.05822 0.4429 -0.08833 0.1399 fixed
count_birth_order3/5 -0.005886 0.05987 -0.09831 -0.1232 0.1115 fixed
count_birth_order4/5 0.03416 0.06337 0.539 -0.09005 0.1584 fixed
count_birth_order5/5 0.06224 0.06476 0.9611 -0.06468 0.1892 fixed
count_birth_order1/5+ -0.03199 0.04473 -0.7153 -0.1197 0.05568 fixed
count_birth_order2/5+ -0.06433 0.0461 -1.395 -0.1547 0.02603 fixed
count_birth_order3/5+ -0.058 0.0452 -1.283 -0.1466 0.03059 fixed
count_birth_order4/5+ -0.03051 0.04429 -0.6888 -0.1173 0.0563 fixed
count_birth_order5/5+ -0.06936 0.04462 -1.554 -0.1568 0.0181 fixed
count_birth_order5+/5+ -0.08347 0.0352 -2.372 -0.1525 -0.01449 fixed
sd_(Intercept).mother_pidlink 0.3767 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8769 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 38051 38134 -19015 38029 NA NA NA
12 38053 38143 -19014 38029 0.3095 1 0.578
16 38058 38179 -19013 38026 2.904 4 0.574
26 38072 38268 -19010 38020 6.247 10 0.7941

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.5229 0.4116 1.27 -0.2838 1.33 fixed
poly(age, 3, raw = TRUE)1 -0.01788 0.04671 -0.3827 -0.1094 0.07368 fixed
poly(age, 3, raw = TRUE)2 0.0007731 0.001667 0.4637 -0.002495 0.004041 fixed
poly(age, 3, raw = TRUE)3 -0.00001439 0.00001888 -0.762 -0.0000514 0.00002262 fixed
male -0.115 0.02403 -4.784 -0.1621 -0.06786 fixed
sibling_count3 0.0267 0.03804 0.7019 -0.04786 0.1013 fixed
sibling_count4 -0.04166 0.04102 -1.016 -0.1221 0.03873 fixed
sibling_count5 -0.06095 0.04688 -1.3 -0.1528 0.03094 fixed
sibling_count5+ -0.09835 0.04118 -2.388 -0.179 -0.01764 fixed
sd_(Intercept).mother_pidlink 0.3583 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8618 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.5171 0.4117 1.256 -0.2897 1.324 fixed
birth_order 0.007062 0.007967 0.8865 -0.008552 0.02268 fixed
poly(age, 3, raw = TRUE)1 -0.01823 0.04672 -0.3902 -0.1098 0.07333 fixed
poly(age, 3, raw = TRUE)2 0.0007744 0.001667 0.4645 -0.002493 0.004042 fixed
poly(age, 3, raw = TRUE)3 -0.00001412 0.00001889 -0.7477 -0.00005114 0.00002289 fixed
male -0.1153 0.02404 -4.796 -0.1624 -0.06818 fixed
sibling_count3 0.02326 0.03824 0.6082 -0.05169 0.09821 fixed
sibling_count4 -0.04983 0.04204 -1.185 -0.1322 0.03258 fixed
sibling_count5 -0.07432 0.04925 -1.509 -0.1709 0.02221 fixed
sibling_count5+ -0.1251 0.05104 -2.451 -0.2251 -0.02505 fixed
sd_(Intercept).mother_pidlink 0.3584 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8618 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.519 0.4125 1.258 -0.2896 1.327 fixed
poly(age, 3, raw = TRUE)1 -0.01781 0.04676 -0.3809 -0.1095 0.07385 fixed
poly(age, 3, raw = TRUE)2 0.0007623 0.001669 0.4567 -0.002509 0.004034 fixed
poly(age, 3, raw = TRUE)3 -0.00001408 0.00001891 -0.7446 -0.00005114 0.00002298 fixed
male -0.115 0.02405 -4.78 -0.1621 -0.06782 fixed
sibling_count3 0.02667 0.03902 0.6834 -0.04981 0.1031 fixed
sibling_count4 -0.04087 0.04367 -0.9359 -0.1265 0.04472 fixed
sibling_count5 -0.05821 0.05172 -1.125 -0.1596 0.04317 fixed
sibling_count5+ -0.1119 0.05254 -2.13 -0.2149 -0.008957 fixed
birth_order_nonlinear2 0.01206 0.031 0.3889 -0.0487 0.07281 fixed
birth_order_nonlinear3 0.0001107 0.03827 0.002894 -0.07489 0.07512 fixed
birth_order_nonlinear4 -0.003256 0.04736 -0.06875 -0.09609 0.08958 fixed
birth_order_nonlinear5 -0.004436 0.05912 -0.07503 -0.1203 0.1114 fixed
birth_order_nonlinear5+ 0.04355 0.05949 0.7321 -0.07304 0.1601 fixed
sd_(Intercept).mother_pidlink 0.3573 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8625 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.5809 0.4133 1.406 -0.2291 1.391 fixed
poly(age, 3, raw = TRUE)1 -0.02288 0.04685 -0.4885 -0.1147 0.06894 fixed
poly(age, 3, raw = TRUE)2 0.0009401 0.001672 0.5622 -0.002338 0.004218 fixed
poly(age, 3, raw = TRUE)3 -0.00001602 0.00001895 -0.8455 -0.00005316 0.00002112 fixed
male -0.115 0.02406 -4.778 -0.1621 -0.06781 fixed
count_birth_order2/2 -0.03987 0.0561 -0.7107 -0.1498 0.07008 fixed
count_birth_order1/3 0.01812 0.04988 0.3633 -0.07964 0.1159 fixed
count_birth_order2/3 0.03543 0.05445 0.6507 -0.07129 0.1422 fixed
count_birth_order3/3 -0.0276 0.06081 -0.4539 -0.1468 0.09158 fixed
count_birth_order1/4 -0.09582 0.06099 -1.571 -0.2154 0.02371 fixed
count_birth_order2/4 0.009069 0.06301 0.1439 -0.1144 0.1326 fixed
count_birth_order3/4 -0.06621 0.06652 -0.9952 -0.1966 0.06418 fixed
count_birth_order4/4 -0.06907 0.06925 -0.9974 -0.2048 0.06666 fixed
count_birth_order1/5 -0.1395 0.08275 -1.686 -0.3017 0.02266 fixed
count_birth_order2/5 -0.03975 0.08852 -0.4491 -0.2132 0.1337 fixed
count_birth_order3/5 -0.08459 0.08298 -1.019 -0.2472 0.07805 fixed
count_birth_order4/5 -0.1223 0.07995 -1.53 -0.279 0.03438 fixed
count_birth_order5/5 0.02383 0.08317 0.2865 -0.1392 0.1868 fixed
count_birth_order1/5+ -0.1065 0.08207 -1.298 -0.2674 0.05433 fixed
count_birth_order2/5+ -0.199 0.08123 -2.45 -0.3582 -0.03983 fixed
count_birth_order3/5+ -0.03194 0.08132 -0.3928 -0.1913 0.1274 fixed
count_birth_order4/5+ -0.08364 0.07662 -1.092 -0.2338 0.06653 fixed
count_birth_order5/5+ -0.2049 0.0724 -2.831 -0.3469 -0.06304 fixed
count_birth_order5+/5+ -0.08616 0.05485 -1.571 -0.1937 0.02135 fixed
sd_(Intercept).mother_pidlink 0.3559 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8628 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 15967 16040 -7972 15945 NA NA NA
12 15968 16048 -7972 15944 0.7869 1 0.375
16 15976 16083 -7972 15944 0.1261 4 0.9981
26 15984 16158 -7966 15932 11.94 10 0.2889

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.5618 0.4105 1.368 -0.2428 1.366 fixed
poly(age, 3, raw = TRUE)1 -0.02052 0.04661 -0.4403 -0.1119 0.07083 fixed
poly(age, 3, raw = TRUE)2 0.0008411 0.001664 0.5055 -0.00242 0.004102 fixed
poly(age, 3, raw = TRUE)3 -0.0000151 0.00001885 -0.801 -0.00005204 0.00002184 fixed
male -0.117 0.02395 -4.887 -0.164 -0.07011 fixed
sibling_count3 0.01493 0.04115 0.3629 -0.06573 0.09559 fixed
sibling_count4 -0.04601 0.04339 -1.061 -0.131 0.03902 fixed
sibling_count5 -0.007199 0.04649 -0.1548 -0.09832 0.08392 fixed
sibling_count5+ -0.08743 0.04073 -2.147 -0.1673 -0.007599 fixed
sd_(Intercept).mother_pidlink 0.3588 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8621 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.5616 0.4106 1.368 -0.2432 1.366 fixed
birth_order 0.000166 0.006964 0.02383 -0.01348 0.01382 fixed
poly(age, 3, raw = TRUE)1 -0.02052 0.04661 -0.4403 -0.1119 0.07084 fixed
poly(age, 3, raw = TRUE)2 0.0008408 0.001664 0.5053 -0.002421 0.004102 fixed
poly(age, 3, raw = TRUE)3 -0.00001509 0.00001885 -0.8003 -0.00005204 0.00002186 fixed
male -0.1171 0.02395 -4.887 -0.164 -0.07011 fixed
sibling_count3 0.01485 0.0413 0.3596 -0.06609 0.09579 fixed
sibling_count4 -0.0462 0.04407 -1.048 -0.1326 0.04018 fixed
sibling_count5 -0.007488 0.04806 -0.1558 -0.1017 0.08671 fixed
sibling_count5+ -0.08804 0.04811 -1.83 -0.1823 0.006261 fixed
sd_(Intercept).mother_pidlink 0.3589 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8622 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.5356 0.4112 1.302 -0.2704 1.342 fixed
poly(age, 3, raw = TRUE)1 -0.01871 0.04664 -0.4011 -0.1101 0.07271 fixed
poly(age, 3, raw = TRUE)2 0.0007845 0.001665 0.4711 -0.002479 0.004048 fixed
poly(age, 3, raw = TRUE)3 -0.00001472 0.00001887 -0.7799 -0.0000517 0.00002227 fixed
male -0.1163 0.02396 -4.853 -0.1632 -0.06932 fixed
sibling_count3 0.01644 0.04204 0.3912 -0.06595 0.09883 fixed
sibling_count4 -0.03607 0.0456 -0.7909 -0.1254 0.05331 fixed
sibling_count5 0.00655 0.05033 0.1302 -0.09209 0.1052 fixed
sibling_count5+ -0.05999 0.04958 -1.21 -0.1572 0.03719 fixed
birth_order_nonlinear2 0.03067 0.03158 0.9712 -0.03123 0.09257 fixed
birth_order_nonlinear3 -0.00388 0.03816 -0.1017 -0.07868 0.07092 fixed
birth_order_nonlinear4 -0.03289 0.04598 -0.7153 -0.123 0.05723 fixed
birth_order_nonlinear5 -0.004899 0.05639 -0.08687 -0.1154 0.1056 fixed
birth_order_nonlinear5+ -0.03593 0.05319 -0.6756 -0.1402 0.06831 fixed
sd_(Intercept).mother_pidlink 0.359 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8622 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.5631 0.4121 1.366 -0.2447 1.371 fixed
poly(age, 3, raw = TRUE)1 -0.02195 0.04672 -0.4699 -0.1135 0.06962 fixed
poly(age, 3, raw = TRUE)2 0.0009012 0.001668 0.5401 -0.002369 0.004171 fixed
poly(age, 3, raw = TRUE)3 -0.00001601 0.00001891 -0.8466 -0.00005307 0.00002105 fixed
male -0.1161 0.02398 -4.841 -0.1631 -0.0691 fixed
count_birth_order2/2 0.031 0.06149 0.5042 -0.08952 0.1515 fixed
count_birth_order1/3 0.03858 0.05405 0.7138 -0.06735 0.1445 fixed
count_birth_order2/3 0.04624 0.05861 0.789 -0.06863 0.1611 fixed
count_birth_order3/3 -0.03029 0.06591 -0.4596 -0.1595 0.09889 fixed
count_birth_order1/4 -0.09284 0.06392 -1.452 -0.2181 0.03244 fixed
count_birth_order2/4 0.04085 0.06489 0.6295 -0.08633 0.168 fixed
count_birth_order3/4 -0.03359 0.07128 -0.4712 -0.1733 0.1061 fixed
count_birth_order4/4 -0.05443 0.0736 -0.7396 -0.1987 0.08983 fixed
count_birth_order1/5 0.01304 0.07582 0.172 -0.1356 0.1616 fixed
count_birth_order2/5 0.04794 0.08135 0.5893 -0.1115 0.2074 fixed
count_birth_order3/5 -0.003747 0.07911 -0.04736 -0.1588 0.1513 fixed
count_birth_order4/5 -0.05214 0.08197 -0.6361 -0.2128 0.1085 fixed
count_birth_order5/5 0.01449 0.0822 0.1763 -0.1466 0.1756 fixed
count_birth_order1/5+ -0.04045 0.07206 -0.5613 -0.1817 0.1008 fixed
count_birth_order2/5+ -0.1055 0.07521 -1.403 -0.2529 0.04191 fixed
count_birth_order3/5+ -0.01007 0.07364 -0.1367 -0.1544 0.1343 fixed
count_birth_order4/5+ -0.08782 0.07098 -1.237 -0.2269 0.05131 fixed
count_birth_order5/5+ -0.07425 0.07291 -1.018 -0.2171 0.06864 fixed
count_birth_order5+/5+ -0.09624 0.05377 -1.79 -0.2016 0.009152 fixed
sd_(Intercept).mother_pidlink 0.3582 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8629 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 16110 16184 -8044 16088 NA NA NA
12 16112 16192 -8044 16088 0.0005688 1 0.981
16 16117 16225 -8043 16085 2.665 4 0.6153
26 16132 16306 -8040 16080 5.799 10 0.8318

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.4531 0.4156 1.09 -0.3615 1.268 fixed
poly(age, 3, raw = TRUE)1 -0.01023 0.04719 -0.2168 -0.1027 0.08226 fixed
poly(age, 3, raw = TRUE)2 0.0004686 0.001685 0.2781 -0.002833 0.003771 fixed
poly(age, 3, raw = TRUE)3 -0.00001034 0.00001909 -0.5417 -0.00004776 0.00002708 fixed
male -0.1113 0.02423 -4.593 -0.1588 -0.0638 fixed
sibling_count3 0.0278 0.03751 0.7411 -0.04572 0.1013 fixed
sibling_count4 -0.03086 0.04071 -0.7581 -0.1107 0.04893 fixed
sibling_count5 -0.05099 0.04803 -1.062 -0.1451 0.04315 fixed
sibling_count5+ -0.1063 0.04162 -2.554 -0.1878 -0.0247 fixed
sd_(Intercept).mother_pidlink 0.3613 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8587 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.4481 0.4157 1.078 -0.3667 1.263 fixed
birth_order 0.007437 0.008192 0.9078 -0.008619 0.02349 fixed
poly(age, 3, raw = TRUE)1 -0.01071 0.04719 -0.2268 -0.1032 0.08179 fixed
poly(age, 3, raw = TRUE)2 0.0004727 0.001685 0.2806 -0.002829 0.003775 fixed
poly(age, 3, raw = TRUE)3 -0.00001009 0.0000191 -0.5283 -0.00004751 0.00002734 fixed
male -0.1115 0.02424 -4.601 -0.159 -0.064 fixed
sibling_count3 0.02415 0.03773 0.6402 -0.04979 0.09809 fixed
sibling_count4 -0.03934 0.04177 -0.9419 -0.1212 0.04253 fixed
sibling_count5 -0.06446 0.05027 -1.282 -0.163 0.03407 fixed
sibling_count5+ -0.1342 0.05175 -2.593 -0.2356 -0.03277 fixed
sd_(Intercept).mother_pidlink 0.3613 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8587 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.4518 0.4165 1.085 -0.3644 1.268 fixed
poly(age, 3, raw = TRUE)1 -0.01037 0.04723 -0.2196 -0.103 0.08221 fixed
poly(age, 3, raw = TRUE)2 0.0004663 0.001686 0.2765 -0.002839 0.003772 fixed
poly(age, 3, raw = TRUE)3 -0.00001016 0.00001911 -0.5316 -0.00004763 0.0000273 fixed
male -0.1112 0.02425 -4.585 -0.1587 -0.06365 fixed
sibling_count3 0.02459 0.03852 0.6384 -0.05091 0.1001 fixed
sibling_count4 -0.03243 0.04344 -0.7466 -0.1176 0.0527 fixed
sibling_count5 -0.04803 0.05255 -0.9139 -0.151 0.05497 fixed
sibling_count5+ -0.119 0.05332 -2.231 -0.2235 -0.01447 fixed
birth_order_nonlinear2 0.009103 0.03086 0.295 -0.05137 0.06958 fixed
birth_order_nonlinear3 0.01424 0.03819 0.3728 -0.06061 0.08909 fixed
birth_order_nonlinear4 -0.008 0.04863 -0.1645 -0.1033 0.08732 fixed
birth_order_nonlinear5 -0.01714 0.06154 -0.2786 -0.1378 0.1035 fixed
birth_order_nonlinear5+ 0.04256 0.06125 0.6948 -0.07749 0.1626 fixed
sd_(Intercept).mother_pidlink 0.3603 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8594 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.4984 0.4175 1.194 -0.3199 1.317 fixed
poly(age, 3, raw = TRUE)1 -0.01432 0.04735 -0.3024 -0.1071 0.07848 fixed
poly(age, 3, raw = TRUE)2 0.0006025 0.001691 0.3564 -0.002711 0.003916 fixed
poly(age, 3, raw = TRUE)3 -0.00001163 0.00001917 -0.6066 -0.00004919 0.00002594 fixed
male -0.1119 0.02427 -4.613 -0.1595 -0.06437 fixed
count_birth_order2/2 -0.02321 0.05452 -0.4258 -0.1301 0.08365 fixed
count_birth_order1/3 0.0164 0.0492 0.3334 -0.08004 0.1128 fixed
count_birth_order2/3 0.0359 0.05434 0.6607 -0.0706 0.1424 fixed
count_birth_order3/3 0.005022 0.05963 0.08422 -0.1119 0.1219 fixed
count_birth_order1/4 -0.05444 0.06119 -0.8898 -0.1744 0.06548 fixed
count_birth_order2/4 0.01122 0.06281 0.1786 -0.1119 0.1343 fixed
count_birth_order3/4 -0.05586 0.06573 -0.8499 -0.1847 0.07297 fixed
count_birth_order4/4 -0.06476 0.06956 -0.9311 -0.2011 0.07157 fixed
count_birth_order1/5 -0.1147 0.08246 -1.39 -0.2763 0.04697 fixed
count_birth_order2/5 -0.0301 0.09102 -0.3307 -0.2085 0.1483 fixed
count_birth_order3/5 -0.04437 0.08684 -0.511 -0.2146 0.1258 fixed
count_birth_order4/5 -0.08581 0.0837 -1.025 -0.2499 0.07823 fixed
count_birth_order5/5 -0.004708 0.08889 -0.05296 -0.1789 0.1695 fixed
count_birth_order1/5+ -0.1188 0.08387 -1.416 -0.2831 0.04562 fixed
count_birth_order2/5+ -0.2097 0.08314 -2.522 -0.3727 -0.04674 fixed
count_birth_order3/5+ -0.02078 0.08233 -0.2523 -0.1821 0.1406 fixed
count_birth_order4/5+ -0.1022 0.0805 -1.27 -0.26 0.05558 fixed
count_birth_order5/5+ -0.1908 0.07401 -2.578 -0.3359 -0.04576 fixed
count_birth_order5+/5+ -0.08773 0.0558 -1.572 -0.1971 0.02165 fixed
sd_(Intercept).mother_pidlink 0.3595 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8599 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 15623 15696 -7801 15601 NA NA NA
12 15624 15704 -7800 15600 0.8254 1 0.3636
16 15632 15739 -7800 15600 0.2888 4 0.9905
26 15645 15818 -7796 15593 7.352 10 0.6918

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Adaptive Numbering

birthorder <- birthorder %>% mutate(outcome = adaptive_numbering)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.3787 0.1518 2.495 0.08118 0.6763 fixed
poly(age, 3, raw = TRUE)1 -0.0186 0.01457 -1.276 -0.04715 0.00996 fixed
poly(age, 3, raw = TRUE)2 0.00058 0.0004296 1.35 -0.000262 0.001422 fixed
poly(age, 3, raw = TRUE)3 -0.00001002 0.000003964 -2.527 -0.00001779 -0.000002248 fixed
male 0.09485 0.01571 6.038 0.06406 0.1256 fixed
sibling_count3 0.032 0.03342 0.9577 -0.03349 0.0975 fixed
sibling_count4 -0.01255 0.03463 -0.3623 -0.08042 0.05533 fixed
sibling_count5 0.0188 0.03622 0.519 -0.05219 0.08979 fixed
sibling_count5+ -0.1474 0.02824 -5.219 -0.2027 -0.09201 fixed
sd_(Intercept).mother_pidlink 0.4629 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8391 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.3788 0.1518 2.495 0.08125 0.6764 fixed
birth_order 0.0008488 0.003398 0.2498 -0.005812 0.007509 fixed
poly(age, 3, raw = TRUE)1 -0.01886 0.01461 -1.291 -0.04749 0.009776 fixed
poly(age, 3, raw = TRUE)2 0.00059 0.0004315 1.367 -0.0002557 0.001436 fixed
poly(age, 3, raw = TRUE)3 -0.00001011 0.000003983 -2.539 -0.00001792 -0.000002306 fixed
male 0.09483 0.01571 6.036 0.06404 0.1256 fixed
sibling_count3 0.03184 0.03343 0.9526 -0.03367 0.09735 fixed
sibling_count4 -0.01309 0.0347 -0.3773 -0.08111 0.05492 fixed
sibling_count5 0.0178 0.03644 0.4886 -0.05361 0.08922 fixed
sibling_count5+ -0.1505 0.03095 -4.863 -0.2112 -0.08985 fixed
sd_(Intercept).mother_pidlink 0.4631 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.839 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.3755 0.1522 2.467 0.07721 0.6739 fixed
poly(age, 3, raw = TRUE)1 -0.01884 0.01461 -1.289 -0.04747 0.0098 fixed
poly(age, 3, raw = TRUE)2 0.000594 0.0004315 1.376 -0.0002518 0.00144 fixed
poly(age, 3, raw = TRUE)3 -0.00001017 0.000003984 -2.554 -0.00001798 -0.000002366 fixed
male 0.0947 0.01571 6.028 0.06391 0.1255 fixed
sibling_count3 0.03351 0.03383 0.9908 -0.03278 0.09981 fixed
sibling_count4 -0.006042 0.03553 -0.1701 -0.07568 0.06359 fixed
sibling_count5 0.01888 0.03763 0.5017 -0.05487 0.09263 fixed
sibling_count5+ -0.1519 0.03232 -4.701 -0.2153 -0.0886 fixed
birth_order_nonlinear2 0.006312 0.02259 0.2794 -0.03796 0.05059 fixed
birth_order_nonlinear3 -0.004016 0.02658 -0.1511 -0.05611 0.04808 fixed
birth_order_nonlinear4 -0.02674 0.03029 -0.8829 -0.08611 0.03263 fixed
birth_order_nonlinear5 0.04054 0.03445 1.177 -0.02699 0.1081 fixed
birth_order_nonlinear5+ 0.0124 0.02911 0.4258 -0.04467 0.06946 fixed
sd_(Intercept).mother_pidlink 0.4633 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.839 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.3813 0.1528 2.495 0.08174 0.6808 fixed
poly(age, 3, raw = TRUE)1 -0.01851 0.01461 -1.267 -0.04714 0.01013 fixed
poly(age, 3, raw = TRUE)2 0.0005781 0.0004316 1.339 -0.0002678 0.001424 fixed
poly(age, 3, raw = TRUE)3 -0.000009971 0.000003985 -2.502 -0.00001778 -0.00000216 fixed
male 0.09384 0.01571 5.974 0.06305 0.1246 fixed
count_birth_order2/2 -0.01282 0.04393 -0.2919 -0.09893 0.07329 fixed
count_birth_order1/3 -0.01112 0.04301 -0.2585 -0.09541 0.07317 fixed
count_birth_order2/3 0.06464 0.04793 1.349 -0.0293 0.1586 fixed
count_birth_order3/3 0.0586 0.05349 1.095 -0.04624 0.1634 fixed
count_birth_order1/4 -0.04414 0.04894 -0.9019 -0.1401 0.05178 fixed
count_birth_order2/4 -0.00802 0.05143 -0.156 -0.1088 0.09277 fixed
count_birth_order3/4 0.01828 0.05567 0.3283 -0.09083 0.1274 fixed
count_birth_order4/4 -0.02113 0.05884 -0.3592 -0.1364 0.09418 fixed
count_birth_order1/5 -0.0386 0.05535 -0.6975 -0.1471 0.06987 fixed
count_birth_order2/5 0.0696 0.05811 1.198 -0.04429 0.1835 fixed
count_birth_order3/5 -0.01428 0.05957 -0.2397 -0.131 0.1025 fixed
count_birth_order4/5 0.002029 0.06299 0.03221 -0.1214 0.1255 fixed
count_birth_order5/5 0.07302 0.06434 1.135 -0.05308 0.1991 fixed
count_birth_order1/5+ -0.0773 0.04455 -1.735 -0.1646 0.01003 fixed
count_birth_order2/5+ -0.1907 0.04587 -4.157 -0.2806 -0.1008 fixed
count_birth_order3/5+ -0.1895 0.04493 -4.217 -0.2775 -0.1014 fixed
count_birth_order4/5+ -0.1979 0.04407 -4.49 -0.2843 -0.1115 fixed
count_birth_order5/5+ -0.1246 0.04435 -2.81 -0.2116 -0.03768 fixed
count_birth_order5+/5+ -0.1466 0.03553 -4.125 -0.2162 -0.07693 fixed
sd_(Intercept).mother_pidlink 0.4635 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8387 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 37830 37913 -18904 37808 NA NA NA
12 37832 37922 -18904 37808 0.0613 1 0.8045
16 37836 37957 -18902 37804 3.594 4 0.4637
26 37840 38036 -18894 37788 15.97 10 0.1004

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5739 0.3823 -1.501 -1.323 0.1753 fixed
poly(age, 3, raw = TRUE)1 0.102 0.04338 2.351 0.01697 0.187 fixed
poly(age, 3, raw = TRUE)2 -0.003674 0.001548 -2.373 -0.006709 -0.0006395 fixed
poly(age, 3, raw = TRUE)3 0.00004194 0.00001754 2.391 0.000007568 0.00007632 fixed
male 0.03722 0.02229 1.67 -0.006472 0.08091 fixed
sibling_count3 -0.02318 0.03623 -0.6397 -0.09419 0.04783 fixed
sibling_count4 -0.07966 0.03922 -2.031 -0.1565 -0.002783 fixed
sibling_count5 -0.09175 0.04495 -2.041 -0.1799 -0.003645 fixed
sibling_count5+ -0.2292 0.03956 -5.795 -0.3067 -0.1517 fixed
sd_(Intercept).mother_pidlink 0.4128 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7721 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.581 0.3823 -1.52 -1.33 0.1683 fixed
birth_order 0.009145 0.007438 1.229 -0.005433 0.02372 fixed
poly(age, 3, raw = TRUE)1 0.1014 0.04338 2.338 0.0164 0.1864 fixed
poly(age, 3, raw = TRUE)2 -0.003667 0.001548 -2.368 -0.006701 -0.0006323 fixed
poly(age, 3, raw = TRUE)3 0.00004223 0.00001754 2.408 0.000007859 0.00007661 fixed
male 0.03681 0.02229 1.651 -0.00688 0.0805 fixed
sibling_count3 -0.02771 0.03642 -0.7608 -0.09909 0.04367 fixed
sibling_count4 -0.09046 0.0402 -2.251 -0.1692 -0.01168 fixed
sibling_count5 -0.1094 0.04718 -2.318 -0.2018 -0.01688 fixed
sibling_count5+ -0.2643 0.04877 -5.419 -0.3599 -0.1687 fixed
sd_(Intercept).mother_pidlink 0.4131 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7719 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5991 0.3831 -1.564 -1.35 0.1518 fixed
poly(age, 3, raw = TRUE)1 0.104 0.04342 2.395 0.0189 0.1891 fixed
poly(age, 3, raw = TRUE)2 -0.003757 0.00155 -2.425 -0.006795 -0.00072 fixed
poly(age, 3, raw = TRUE)3 0.00004323 0.00001756 2.462 0.000008815 0.00007764 fixed
male 0.03724 0.0223 1.67 -0.006459 0.08094 fixed
sibling_count3 -0.03169 0.0371 -0.8543 -0.1044 0.04102 fixed
sibling_count4 -0.09277 0.04163 -2.228 -0.1744 -0.01117 fixed
sibling_count5 -0.1259 0.04936 -2.55 -0.2226 -0.02913 fixed
sibling_count5+ -0.2684 0.05009 -5.359 -0.3666 -0.1703 fixed
birth_order_nonlinear2 0.02271 0.02845 0.7982 -0.03305 0.07847 fixed
birth_order_nonlinear3 0.0366 0.03519 1.04 -0.03238 0.1056 fixed
birth_order_nonlinear4 0.02322 0.04359 0.5326 -0.06222 0.1087 fixed
birth_order_nonlinear5 0.1172 0.05437 2.156 0.01065 0.2238 fixed
birth_order_nonlinear5+ 0.04245 0.05526 0.7682 -0.06586 0.1508 fixed
sd_(Intercept).mother_pidlink 0.413 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.772 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5894 0.384 -1.535 -1.342 0.1632 fixed
poly(age, 3, raw = TRUE)1 0.1022 0.04351 2.348 0.01689 0.1875 fixed
poly(age, 3, raw = TRUE)2 -0.003663 0.001553 -2.358 -0.006707 -0.0006186 fixed
poly(age, 3, raw = TRUE)3 0.00004183 0.0000176 2.376 0.000007331 0.00007633 fixed
male 0.03646 0.02231 1.634 -0.007272 0.08019 fixed
count_birth_order2/2 0.02404 0.05166 0.4654 -0.07721 0.1253 fixed
count_birth_order1/3 -0.05451 0.04671 -1.167 -0.1461 0.03704 fixed
count_birth_order2/3 0.01584 0.05088 0.3112 -0.0839 0.1156 fixed
count_birth_order3/3 0.01729 0.05675 0.3047 -0.09393 0.1285 fixed
count_birth_order1/4 -0.1283 0.05698 -2.251 -0.2399 -0.0166 fixed
count_birth_order2/4 -0.06935 0.05883 -1.179 -0.1847 0.04596 fixed
count_birth_order3/4 -0.03725 0.06203 -0.6005 -0.1588 0.08433 fixed
count_birth_order4/4 -0.03674 0.06447 -0.5699 -0.1631 0.08962 fixed
count_birth_order1/5 -0.03692 0.07709 -0.4789 -0.188 0.1142 fixed
count_birth_order2/5 -0.1199 0.0823 -1.457 -0.2812 0.04141 fixed
count_birth_order3/5 -0.154 0.07716 -1.996 -0.3052 -0.002782 fixed
count_birth_order4/5 -0.1267 0.0744 -1.704 -0.2726 0.01907 fixed
count_birth_order5/5 0.002534 0.07733 0.03277 -0.149 0.1541 fixed
count_birth_order1/5+ -0.1809 0.07635 -2.37 -0.3306 -0.03129 fixed
count_birth_order2/5+ -0.3047 0.07551 -4.035 -0.4527 -0.1567 fixed
count_birth_order3/5+ -0.2237 0.07552 -2.963 -0.3718 -0.07573 fixed
count_birth_order4/5+ -0.2646 0.07115 -3.718 -0.404 -0.1251 fixed
count_birth_order5/5+ -0.1598 0.06737 -2.372 -0.2919 -0.02778 fixed
count_birth_order5+/5+ -0.2265 0.05166 -4.384 -0.3277 -0.1252 fixed
sd_(Intercept).mother_pidlink 0.413 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7722 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 15107 15181 -7543 15085 NA NA NA
12 15108 15188 -7542 15084 1.512 1 0.2188
16 15112 15219 -7540 15080 3.495 4 0.4787
26 15124 15298 -7536 15072 7.814 10 0.647

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.547 0.381 -1.436 -1.294 0.1998 fixed
poly(age, 3, raw = TRUE)1 0.09819 0.04326 2.27 0.01341 0.183 fixed
poly(age, 3, raw = TRUE)2 -0.003581 0.001544 -2.319 -0.006608 -0.0005544 fixed
poly(age, 3, raw = TRUE)3 0.0000409 0.0000175 2.338 0.00000661 0.00007519 fixed
male 0.03865 0.0222 1.741 -0.004853 0.08215 fixed
sibling_count3 -0.001308 0.03913 -0.03344 -0.078 0.07538 fixed
sibling_count4 -0.05356 0.04142 -1.293 -0.1347 0.02763 fixed
sibling_count5 -0.04134 0.04448 -0.9294 -0.1285 0.04584 fixed
sibling_count5+ -0.1437 0.03893 -3.692 -0.22 -0.06743 fixed
sd_(Intercept).mother_pidlink 0.4147 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7712 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.548 0.3811 -1.438 -1.295 0.1989 fixed
birth_order 0.0009386 0.006519 0.144 -0.01184 0.01372 fixed
poly(age, 3, raw = TRUE)1 0.09817 0.04326 2.269 0.01338 0.183 fixed
poly(age, 3, raw = TRUE)2 -0.003582 0.001544 -2.319 -0.006608 -0.0005547 fixed
poly(age, 3, raw = TRUE)3 0.00004095 0.0000175 2.34 0.000006647 0.00007524 fixed
male 0.03861 0.0222 1.739 -0.004894 0.08212 fixed
sibling_count3 -0.001773 0.03926 -0.04515 -0.07873 0.07518 fixed
sibling_count4 -0.05462 0.04208 -1.298 -0.1371 0.02785 fixed
sibling_count5 -0.04301 0.04597 -0.9356 -0.1331 0.04709 fixed
sibling_count5+ -0.1472 0.04585 -3.211 -0.2371 -0.05736 fixed
sd_(Intercept).mother_pidlink 0.4149 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7712 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5634 0.3817 -1.476 -1.311 0.1846 fixed
poly(age, 3, raw = TRUE)1 0.09955 0.04328 2.3 0.01473 0.1844 fixed
poly(age, 3, raw = TRUE)2 -0.003632 0.001545 -2.35 -0.00666 -0.0006033 fixed
poly(age, 3, raw = TRUE)3 0.00004154 0.00001751 2.373 0.000007224 0.00007586 fixed
male 0.03953 0.0222 1.781 -0.003975 0.08303 fixed
sibling_count3 0.0002314 0.03989 0.005801 -0.07796 0.07842 fixed
sibling_count4 -0.04784 0.04341 -1.102 -0.1329 0.03725 fixed
sibling_count5 -0.05674 0.04795 -1.183 -0.1507 0.03724 fixed
sibling_count5+ -0.1505 0.04712 -3.194 -0.2429 -0.05815 fixed
birth_order_nonlinear2 0.01206 0.02899 0.416 -0.04476 0.06888 fixed
birth_order_nonlinear3 -0.007123 0.03509 -0.203 -0.0759 0.06165 fixed
birth_order_nonlinear4 -0.02017 0.04232 -0.4766 -0.1031 0.06278 fixed
birth_order_nonlinear5 0.119 0.05185 2.294 0.01734 0.2206 fixed
birth_order_nonlinear5+ -0.007059 0.04951 -0.1426 -0.1041 0.08997 fixed
sd_(Intercept).mother_pidlink 0.4146 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.771 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5329 0.3822 -1.394 -1.282 0.2162 fixed
poly(age, 3, raw = TRUE)1 0.09492 0.04332 2.191 0.01003 0.1798 fixed
poly(age, 3, raw = TRUE)2 -0.00344 0.001547 -2.224 -0.006472 -0.0004087 fixed
poly(age, 3, raw = TRUE)3 0.00003911 0.00001753 2.231 0.000004748 0.00007347 fixed
male 0.03859 0.0222 1.739 -0.004913 0.0821 fixed
count_birth_order2/2 0.02549 0.05653 0.4508 -0.08532 0.1363 fixed
count_birth_order1/3 -0.01494 0.05051 -0.2957 -0.1139 0.08406 fixed
count_birth_order2/3 0.0181 0.05465 0.3312 -0.08901 0.1252 fixed
count_birth_order3/3 0.03519 0.06135 0.5737 -0.08505 0.1554 fixed
count_birth_order1/4 -0.1124 0.05962 -1.886 -0.2293 0.004432 fixed
count_birth_order2/4 0.02033 0.06071 0.3348 -0.09866 0.1393 fixed
count_birth_order3/4 -0.01424 0.06636 -0.2146 -0.1443 0.1158 fixed
count_birth_order4/4 -0.06868 0.06842 -1.004 -0.2028 0.06542 fixed
count_birth_order1/5 0.01101 0.07059 0.1559 -0.1274 0.1494 fixed
count_birth_order2/5 -0.05006 0.07557 -0.6624 -0.1982 0.09805 fixed
count_birth_order3/5 -0.1674 0.07348 -2.278 -0.3114 -0.02335 fixed
count_birth_order4/5 -0.07244 0.07609 -0.952 -0.2216 0.07669 fixed
count_birth_order5/5 0.1097 0.0763 1.438 -0.03981 0.2593 fixed
count_birth_order1/5+ -0.0476 0.06705 -0.7099 -0.179 0.08382 fixed
count_birth_order2/5+ -0.2193 0.06985 -3.139 -0.3561 -0.08236 fixed
count_birth_order3/5+ -0.1526 0.06844 -2.229 -0.2867 -0.01843 fixed
count_birth_order4/5+ -0.1628 0.06588 -2.471 -0.2919 -0.03363 fixed
count_birth_order5/5+ -0.06021 0.06758 -0.891 -0.1927 0.07224 fixed
count_birth_order5+/5+ -0.1548 0.05053 -3.063 -0.2538 -0.05575 fixed
sd_(Intercept).mother_pidlink 0.4146 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7706 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 15237 15311 -7608 15215 NA NA NA
12 15239 15319 -7608 15215 0.02042 1 0.8864
16 15239 15346 -7604 15207 7.892 4 0.09562
26 15244 15418 -7596 15192 15.53 10 0.1138

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.6109 0.3868 -1.579 -1.369 0.1473 fixed
poly(age, 3, raw = TRUE)1 0.1055 0.04391 2.404 0.01948 0.1916 fixed
poly(age, 3, raw = TRUE)2 -0.003789 0.001568 -2.417 -0.006861 -0.0007159 fixed
poly(age, 3, raw = TRUE)3 0.00004303 0.00001777 2.422 0.000008203 0.00007785 fixed
male 0.03418 0.02252 1.518 -0.009957 0.07831 fixed
sibling_count3 -0.01054 0.03585 -0.294 -0.08081 0.05973 fixed
sibling_count4 -0.08272 0.03907 -2.117 -0.1593 -0.006154 fixed
sibling_count5 -0.07539 0.04627 -1.629 -0.1661 0.01529 fixed
sibling_count5+ -0.2279 0.04017 -5.672 -0.3066 -0.1491 fixed
sd_(Intercept).mother_pidlink 0.4197 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7695 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.6191 0.3868 -1.601 -1.377 0.1389 fixed
birth_order 0.0126 0.007659 1.645 -0.002409 0.02762 fixed
poly(age, 3, raw = TRUE)1 0.1046 0.04391 2.383 0.01855 0.1907 fixed
poly(age, 3, raw = TRUE)2 -0.003775 0.001567 -2.408 -0.006847 -0.0007029 fixed
poly(age, 3, raw = TRUE)3 0.0000434 0.00001777 2.443 0.00000858 0.00007822 fixed
male 0.03386 0.02252 1.504 -0.01027 0.07799 fixed
sibling_count3 -0.01685 0.03605 -0.4673 -0.08751 0.05382 fixed
sibling_count4 -0.0974 0.04007 -2.431 -0.1759 -0.01887 fixed
sibling_count5 -0.09864 0.04838 -2.039 -0.1935 -0.003825 fixed
sibling_count5+ -0.2758 0.04961 -5.559 -0.373 -0.1785 fixed
sd_(Intercept).mother_pidlink 0.42 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7693 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.638 0.3876 -1.646 -1.398 0.1216 fixed
poly(age, 3, raw = TRUE)1 0.1074 0.04395 2.444 0.02129 0.1936 fixed
poly(age, 3, raw = TRUE)2 -0.003874 0.001569 -2.469 -0.006949 -0.0007993 fixed
poly(age, 3, raw = TRUE)3 0.00004448 0.00001778 2.501 0.000009626 0.00007934 fixed
male 0.03404 0.02252 1.511 -0.0101 0.07818 fixed
sibling_count3 -0.0171 0.03674 -0.4653 -0.08911 0.05492 fixed
sibling_count4 -0.09868 0.04153 -2.376 -0.1801 -0.01729 fixed
sibling_count5 -0.1131 0.05037 -2.245 -0.2118 -0.01436 fixed
sibling_count5+ -0.2769 0.05098 -5.432 -0.3768 -0.177 fixed
birth_order_nonlinear2 0.03343 0.02836 1.179 -0.02215 0.08901 fixed
birth_order_nonlinear3 0.02907 0.03517 0.8265 -0.03986 0.09799 fixed
birth_order_nonlinear4 0.04795 0.04482 1.07 -0.03988 0.1358 fixed
birth_order_nonlinear5 0.1317 0.05668 2.324 0.02065 0.2428 fixed
birth_order_nonlinear5+ 0.06088 0.05697 1.069 -0.05078 0.1725 fixed
sd_(Intercept).mother_pidlink 0.4191 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7697 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.6225 0.3884 -1.602 -1.384 0.1389 fixed
poly(age, 3, raw = TRUE)1 0.1047 0.04404 2.378 0.01842 0.1911 fixed
poly(age, 3, raw = TRUE)2 -0.003746 0.001573 -2.382 -0.006828 -0.0006634 fixed
poly(age, 3, raw = TRUE)3 0.00004268 0.00001783 2.394 0.000007732 0.00007762 fixed
male 0.03277 0.02253 1.454 -0.0114 0.07693 fixed
count_birth_order2/2 0.0373 0.05023 0.7426 -0.06115 0.1358 fixed
count_birth_order1/3 -0.0397 0.04617 -0.86 -0.1302 0.05078 fixed
count_birth_order2/3 0.04364 0.05085 0.8583 -0.05602 0.1433 fixed
count_birth_order3/3 0.02499 0.05573 0.4484 -0.08424 0.1342 fixed
count_birth_order1/4 -0.1454 0.05724 -2.54 -0.2576 -0.03319 fixed
count_birth_order2/4 -0.06066 0.05873 -1.033 -0.1758 0.05444 fixed
count_birth_order3/4 -0.04567 0.06138 -0.744 -0.166 0.07463 fixed
count_birth_order4/4 -0.009882 0.06483 -0.1524 -0.1369 0.1172 fixed
count_birth_order1/5 -0.03364 0.07699 -0.4369 -0.1845 0.1173 fixed
count_birth_order2/5 -0.1226 0.08473 -1.447 -0.2887 0.04345 fixed
count_birth_order3/5 -0.1298 0.08084 -1.606 -0.2883 0.02861 fixed
count_birth_order4/5 -0.08387 0.07798 -1.075 -0.2367 0.06897 fixed
count_birth_order5/5 0.0319 0.08276 0.3854 -0.1303 0.1941 fixed
count_birth_order1/5+ -0.1417 0.07812 -1.814 -0.2948 0.01137 fixed
count_birth_order2/5+ -0.3013 0.07739 -3.893 -0.4529 -0.1496 fixed
count_birth_order3/5+ -0.2695 0.07655 -3.521 -0.4196 -0.1195 fixed
count_birth_order4/5+ -0.2658 0.07478 -3.554 -0.4123 -0.1192 fixed
count_birth_order5/5+ -0.1536 0.06895 -2.227 -0.2887 -0.01844 fixed
count_birth_order5+/5+ -0.2161 0.05268 -4.102 -0.3193 -0.1128 fixed
sd_(Intercept).mother_pidlink 0.4191 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.7697 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 14808 14882 -7393 14786 NA NA NA
12 14808 14888 -7392 14784 2.71 1 0.09974
16 14813 14919 -7390 14781 3.147 4 0.5336
26 14822 14995 -7385 14770 10.48 10 0.3993

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Personality

Extraversion

birthorder <- birthorder %>% mutate(outcome = big5_ext)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.5858 0.1594 3.676 0.2735 0.8982 fixed
poly(age, 3, raw = TRUE)1 -0.0431 0.01526 -2.824 -0.07301 -0.01319 fixed
poly(age, 3, raw = TRUE)2 0.001286 0.0004482 2.868 0.0004071 0.002164 fixed
poly(age, 3, raw = TRUE)3 -0.00001265 0.000004122 -3.068 -0.00002073 -0.000004568 fixed
male -0.2258 0.01689 -13.37 -0.2589 -0.1927 fixed
sibling_count3 0.009322 0.03332 0.2798 -0.05598 0.07462 fixed
sibling_count4 -0.02026 0.03419 -0.5925 -0.08728 0.04676 fixed
sibling_count5 -0.02693 0.03538 -0.7611 -0.09627 0.04241 fixed
sibling_count5+ -0.01081 0.02791 -0.3874 -0.06552 0.0439 fixed
sd_(Intercept).mother_pidlink 0.2287 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9711 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.5841 0.1594 3.665 0.2718 0.8965 fixed
birth_order -0.003173 0.003493 -0.9085 -0.01002 0.003673 fixed
poly(age, 3, raw = TRUE)1 -0.04216 0.01529 -2.757 -0.07214 -0.01219 fixed
poly(age, 3, raw = TRUE)2 0.001253 0.0004497 2.786 0.0003715 0.002134 fixed
poly(age, 3, raw = TRUE)3 -0.00001235 0.000004135 -2.987 -0.00002045 -0.000004247 fixed
male -0.2257 0.01689 -13.36 -0.2588 -0.1926 fixed
sibling_count3 0.01013 0.03333 0.304 -0.05519 0.07545 fixed
sibling_count4 -0.01798 0.03428 -0.5243 -0.08517 0.04922 fixed
sibling_count5 -0.02301 0.03564 -0.6457 -0.09286 0.04684 fixed
sibling_count5+ 0.001366 0.03097 0.0441 -0.05933 0.06206 fixed
sd_(Intercept).mother_pidlink 0.2281 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9712 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.5626 0.1598 3.521 0.2494 0.8758 fixed
poly(age, 3, raw = TRUE)1 -0.04263 0.0153 -2.787 -0.07262 -0.01265 fixed
poly(age, 3, raw = TRUE)2 0.001264 0.0004498 2.809 0.000382 0.002145 fixed
poly(age, 3, raw = TRUE)3 -0.00001237 0.000004136 -2.991 -0.00002048 -0.000004262 fixed
male -0.2257 0.01689 -13.36 -0.2588 -0.1926 fixed
sibling_count3 0.01351 0.03385 0.3991 -0.05284 0.07985 fixed
sibling_count4 -0.01763 0.03537 -0.4984 -0.08695 0.05169 fixed
sibling_count5 -0.02496 0.03719 -0.6712 -0.09785 0.04793 fixed
sibling_count5+ -0.005669 0.03271 -0.1733 -0.06978 0.05845 fixed
birth_order_nonlinear2 0.05536 0.02472 2.239 0.006905 0.1038 fixed
birth_order_nonlinear3 -0.003555 0.02919 -0.1218 -0.06076 0.05365 fixed
birth_order_nonlinear4 0.02086 0.03323 0.6278 -0.04427 0.086 fixed
birth_order_nonlinear5 0.02606 0.03779 0.6897 -0.048 0.1001 fixed
birth_order_nonlinear5+ 0.01193 0.031 0.3846 -0.04884 0.07269 fixed
sd_(Intercept).mother_pidlink 0.229 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9709 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.5658 0.1605 3.525 0.2511 0.8804 fixed
poly(age, 3, raw = TRUE)1 -0.04235 0.0153 -2.767 -0.07234 -0.01235 fixed
poly(age, 3, raw = TRUE)2 0.001259 0.0004499 2.799 0.0003773 0.002141 fixed
poly(age, 3, raw = TRUE)3 -0.00001236 0.000004137 -2.987 -0.00002047 -0.00000425 fixed
male -0.2255 0.0169 -13.35 -0.2587 -0.1924 fixed
count_birth_order2/2 0.03555 0.04808 0.7393 -0.05869 0.1298 fixed
count_birth_order1/3 0.006548 0.04502 0.1454 -0.08169 0.09478 fixed
count_birth_order2/3 0.0579 0.05035 1.15 -0.04078 0.1566 fixed
count_birth_order3/3 0.006575 0.05641 0.1166 -0.104 0.1171 fixed
count_birth_order1/4 -0.03226 0.05142 -0.6274 -0.133 0.06852 fixed
count_birth_order2/4 0.09717 0.05408 1.797 -0.008825 0.2032 fixed
count_birth_order3/4 -0.06283 0.05877 -1.069 -0.178 0.05235 fixed
count_birth_order4/4 -0.0513 0.06223 -0.8244 -0.1733 0.07066 fixed
count_birth_order1/5 -0.04349 0.05836 -0.7453 -0.1579 0.07088 fixed
count_birth_order2/5 -0.01869 0.06139 -0.3045 -0.139 0.1016 fixed
count_birth_order3/5 -0.0007656 0.06302 -0.01215 -0.1243 0.1227 fixed
count_birth_order4/5 0.04999 0.06676 0.7488 -0.08086 0.1808 fixed
count_birth_order5/5 -0.04369 0.06825 -0.6402 -0.1775 0.09008 fixed
count_birth_order1/5+ -0.01372 0.04714 -0.2911 -0.1061 0.07867 fixed
count_birth_order2/5+ 0.03022 0.04863 0.6215 -0.06509 0.1255 fixed
count_birth_order3/5+ -0.01704 0.04765 -0.3577 -0.1104 0.07634 fixed
count_birth_order4/5+ 0.005439 0.04669 0.1165 -0.08607 0.09695 fixed
count_birth_order5/5+ 0.02545 0.04704 0.5409 -0.06675 0.1176 fixed
count_birth_order5+/5+ -0.001348 0.03641 -0.03702 -0.07272 0.07002 fixed
sd_(Intercept).mother_pidlink 0.2285 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9712 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 39460 39543 -19719 39438 NA NA NA
12 39461 39552 -19719 39437 0.8272 1 0.3631
16 39464 39585 -19716 39432 5.522 4 0.2378
26 39477 39673 -19713 39425 6.545 10 0.7676

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.152 0.4495 2.564 0.2714 2.033 fixed
poly(age, 3, raw = TRUE)1 -0.1016 0.05098 -1.992 -0.2015 -0.001649 fixed
poly(age, 3, raw = TRUE)2 0.003476 0.001819 1.912 -0.00008788 0.007041 fixed
poly(age, 3, raw = TRUE)3 -0.00003723 0.00002058 -1.809 -0.00007757 0.00000311 fixed
male -0.2631 0.02631 -10 -0.3147 -0.2115 fixed
sibling_count3 -0.03372 0.04026 -0.8374 -0.1126 0.0452 fixed
sibling_count4 -0.08771 0.04309 -2.036 -0.1722 -0.003257 fixed
sibling_count5 -0.08529 0.04889 -1.744 -0.1811 0.01054 fixed
sibling_count5+ -0.09583 0.04277 -2.241 -0.1797 -0.012 fixed
sd_(Intercept).mother_pidlink 0.1894 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9934 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.141 0.4495 2.539 0.2601 2.022 fixed
birth_order 0.01356 0.008594 1.578 -0.003286 0.0304 fixed
poly(age, 3, raw = TRUE)1 -0.1021 0.05098 -2.003 -0.202 -0.002184 fixed
poly(age, 3, raw = TRUE)2 0.003471 0.001818 1.909 -0.00009333 0.007034 fixed
poly(age, 3, raw = TRUE)3 -0.00003665 0.00002058 -1.78 -0.00007699 0.000003697 fixed
male -0.2636 0.02631 -10.02 -0.3152 -0.2121 fixed
sibling_count3 -0.04022 0.04048 -0.9935 -0.1196 0.03912 fixed
sibling_count4 -0.103 0.04417 -2.331 -0.1895 -0.01639 fixed
sibling_count5 -0.1104 0.05143 -2.146 -0.2112 -0.009566 fixed
sibling_count5+ -0.1462 0.05344 -2.736 -0.251 -0.04149 fixed
sd_(Intercept).mother_pidlink 0.1924 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9927 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.11 0.4497 2.467 0.228 1.991 fixed
poly(age, 3, raw = TRUE)1 -0.09789 0.05096 -1.921 -0.1978 0.001998 fixed
poly(age, 3, raw = TRUE)2 0.003318 0.001818 1.825 -0.0002447 0.006882 fixed
poly(age, 3, raw = TRUE)3 -0.00003493 0.00002058 -1.697 -0.00007527 0.000005413 fixed
male -0.2635 0.02629 -10.02 -0.315 -0.2119 fixed
sibling_count3 -0.04151 0.04136 -1.004 -0.1226 0.03955 fixed
sibling_count4 -0.1168 0.04604 -2.538 -0.2071 -0.0266 fixed
sibling_count5 -0.1589 0.0543 -2.927 -0.2653 -0.05248 fixed
sibling_count5+ -0.1572 0.05516 -2.85 -0.2653 -0.04908 fixed
birth_order_nonlinear2 0.03992 0.03448 1.158 -0.02766 0.1075 fixed
birth_order_nonlinear3 0.03437 0.04242 0.8102 -0.04877 0.1175 fixed
birth_order_nonlinear4 0.1099 0.05233 2.101 0.007368 0.2125 fixed
birth_order_nonlinear5 0.2364 0.06542 3.614 0.1082 0.3647 fixed
birth_order_nonlinear5+ 0.03262 0.06472 0.504 -0.09422 0.1595 fixed
sd_(Intercept).mother_pidlink 0.1872 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9929 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.085 0.4508 2.407 0.2013 1.968 fixed
poly(age, 3, raw = TRUE)1 -0.09453 0.05108 -1.851 -0.1946 0.005591 fixed
poly(age, 3, raw = TRUE)2 0.003205 0.001822 1.758 -0.0003673 0.006776 fixed
poly(age, 3, raw = TRUE)3 -0.00003372 0.00002064 -1.634 -0.00007416 0.000006732 fixed
male -0.2623 0.02632 -9.964 -0.3138 -0.2107 fixed
count_birth_order2/2 0.02096 0.06225 0.3367 -0.1011 0.143 fixed
count_birth_order1/3 -0.08328 0.05417 -1.537 -0.1895 0.02289 fixed
count_birth_order2/3 0.03793 0.05923 0.6403 -0.07817 0.154 fixed
count_birth_order3/3 -0.01001 0.06626 -0.151 -0.1399 0.1199 fixed
count_birth_order1/4 -0.06131 0.06625 -0.9255 -0.1912 0.06854 fixed
count_birth_order2/4 -0.1186 0.06863 -1.728 -0.2531 0.01589 fixed
count_birth_order3/4 -0.1046 0.07259 -1.442 -0.2469 0.03763 fixed
count_birth_order4/4 -0.04082 0.07548 -0.5408 -0.1887 0.1071 fixed
count_birth_order1/5 -0.1233 0.09008 -1.368 -0.2998 0.05329 fixed
count_birth_order2/5 -0.163 0.09652 -1.689 -0.3522 0.02615 fixed
count_birth_order3/5 -0.1296 0.0905 -1.432 -0.307 0.04775 fixed
count_birth_order4/5 -0.05215 0.08716 -0.5983 -0.223 0.1187 fixed
count_birth_order5/5 0.05478 0.09073 0.6037 -0.1231 0.2326 fixed
count_birth_order1/5+ -0.2236 0.08934 -2.503 -0.3987 -0.04849 fixed
count_birth_order2/5+ -0.1293 0.0885 -1.461 -0.3028 0.04414 fixed
count_birth_order3/5+ -0.1131 0.08873 -1.275 -0.287 0.06082 fixed
count_birth_order4/5+ -0.0233 0.08363 -0.2786 -0.1872 0.1406 fixed
count_birth_order5/5+ 0.08399 0.07903 1.063 -0.07091 0.2389 fixed
count_birth_order5+/5+ -0.1312 0.05874 -2.234 -0.2463 -0.01608 fixed
sd_(Intercept).mother_pidlink 0.1883 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9931 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 16954 17028 -8466 16932 NA NA NA
12 16954 17034 -8465 16930 2.483 1 0.1151
16 16948 17055 -8458 16916 13.26 4 0.01007
26 16963 17137 -8456 16911 5.148 10 0.8811

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.163 0.4489 2.591 0.2832 2.043 fixed
poly(age, 3, raw = TRUE)1 -0.1022 0.05094 -2.007 -0.2021 -0.002389 fixed
poly(age, 3, raw = TRUE)2 0.003524 0.001817 1.939 -0.00003806 0.007086 fixed
poly(age, 3, raw = TRUE)3 -0.00003799 0.00002057 -1.847 -0.00007832 0.000002332 fixed
male -0.2654 0.02625 -10.11 -0.3169 -0.214 fixed
sibling_count3 -0.03632 0.04368 -0.8315 -0.1219 0.04929 fixed
sibling_count4 -0.07942 0.04582 -1.733 -0.1692 0.01039 fixed
sibling_count5 -0.1033 0.04875 -2.119 -0.1988 -0.007751 fixed
sibling_count5+ -0.1059 0.04277 -2.477 -0.1898 -0.02212 fixed
sd_(Intercept).mother_pidlink 0.1912 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9951 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.15 0.4489 2.561 0.2698 2.029 fixed
birth_order 0.01209 0.007482 1.616 -0.002573 0.02676 fixed
poly(age, 3, raw = TRUE)1 -0.1023 0.05093 -2.009 -0.2021 -0.002473 fixed
poly(age, 3, raw = TRUE)2 0.003505 0.001817 1.929 -0.00005715 0.007066 fixed
poly(age, 3, raw = TRUE)3 -0.00003732 0.00002058 -1.813 -0.00007765 0.000003014 fixed
male -0.2658 0.02625 -10.13 -0.3173 -0.2144 fixed
sibling_count3 -0.04207 0.04384 -0.9598 -0.128 0.04385 fixed
sibling_count4 -0.0925 0.04655 -1.987 -0.1837 -0.001275 fixed
sibling_count5 -0.1239 0.05041 -2.458 -0.2227 -0.0251 fixed
sibling_count5+ -0.1499 0.05073 -2.954 -0.2493 -0.05044 fixed
sd_(Intercept).mother_pidlink 0.1956 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9941 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.141 0.4493 2.539 0.26 2.021 fixed
poly(age, 3, raw = TRUE)1 -0.1002 0.05094 -1.968 -0.2001 -0.0003941 fixed
poly(age, 3, raw = TRUE)2 0.003429 0.001818 1.887 -0.0001333 0.006992 fixed
poly(age, 3, raw = TRUE)3 -0.00003645 0.00002059 -1.771 -0.00007679 0.0000039 fixed
male -0.2649 0.02625 -10.09 -0.3164 -0.2135 fixed
sibling_count3 -0.04594 0.0447 -1.028 -0.1335 0.04167 fixed
sibling_count4 -0.09413 0.04833 -1.948 -0.1889 0.0005931 fixed
sibling_count5 -0.1497 0.05306 -2.821 -0.2537 -0.04569 fixed
sibling_count5+ -0.1589 0.05247 -3.027 -0.2617 -0.05601 fixed
birth_order_nonlinear2 0.02093 0.03515 0.5955 -0.04796 0.08982 fixed
birth_order_nonlinear3 0.04145 0.04233 0.9792 -0.04151 0.1244 fixed
birth_order_nonlinear4 0.02811 0.05083 0.5529 -0.07152 0.1277 fixed
birth_order_nonlinear5 0.1845 0.06237 2.959 0.06228 0.3068 fixed
birth_order_nonlinear5+ 0.06251 0.0578 1.081 -0.05078 0.1758 fixed
sd_(Intercept).mother_pidlink 0.1939 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9942 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.137 0.4502 2.525 0.2545 2.019 fixed
poly(age, 3, raw = TRUE)1 -0.09952 0.05102 -1.95 -0.1995 0.0004835 fixed
poly(age, 3, raw = TRUE)2 0.003409 0.001821 1.872 -0.0001598 0.006978 fixed
poly(age, 3, raw = TRUE)3 -0.00003628 0.00002062 -1.759 -0.0000767 0.000004144 fixed
male -0.265 0.02627 -10.08 -0.3165 -0.2135 fixed
count_birth_order2/2 0.01043 0.06822 0.153 -0.1233 0.1441 fixed
count_birth_order1/3 -0.08321 0.05879 -1.415 -0.1984 0.03202 fixed
count_birth_order2/3 0.01868 0.06381 0.2928 -0.1064 0.1437 fixed
count_birth_order3/3 -0.0109 0.07186 -0.1517 -0.1517 0.1299 fixed
count_birth_order1/4 -0.05074 0.06955 -0.7296 -0.1871 0.08557 fixed
count_birth_order2/4 -0.06652 0.07091 -0.938 -0.2055 0.07247 fixed
count_birth_order3/4 -0.06716 0.07791 -0.862 -0.2199 0.08554 fixed
count_birth_order4/4 -0.1452 0.08033 -1.808 -0.3026 0.01224 fixed
count_birth_order1/5 -0.1287 0.08259 -1.559 -0.2906 0.03312 fixed
count_birth_order2/5 -0.1947 0.08878 -2.193 -0.3687 -0.02069 fixed
count_birth_order3/5 -0.1136 0.08636 -1.316 -0.2829 0.05564 fixed
count_birth_order4/5 -0.123 0.08956 -1.373 -0.2985 0.05257 fixed
count_birth_order5/5 0.06422 0.08981 0.7151 -0.1118 0.2403 fixed
count_birth_order1/5+ -0.1769 0.07848 -2.255 -0.3308 -0.02313 fixed
count_birth_order2/5+ -0.1904 0.08206 -2.32 -0.3512 -0.02953 fixed
count_birth_order3/5+ -0.1052 0.08055 -1.306 -0.2631 0.05268 fixed
count_birth_order4/5+ -0.07007 0.07755 -0.9036 -0.2221 0.08192 fixed
count_birth_order5/5+ -0.001767 0.07974 -0.02216 -0.1581 0.1545 fixed
count_birth_order5+/5+ -0.1002 0.05776 -1.736 -0.2135 0.01295 fixed
sd_(Intercept).mother_pidlink 0.1906 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9952 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 17123 17197 -8551 17101 NA NA NA
12 17123 17203 -8549 17099 2.6 1 0.1069
16 17124 17231 -8546 17092 6.411 4 0.1705
26 17138 17312 -8543 17086 6.211 10 0.7972

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.072 0.4545 2.359 0.1814 1.963 fixed
poly(age, 3, raw = TRUE)1 -0.0908 0.05157 -1.761 -0.1919 0.01027 fixed
poly(age, 3, raw = TRUE)2 0.003071 0.00184 1.669 -0.0005349 0.006678 fixed
poly(age, 3, raw = TRUE)3 -0.00003239 0.00002084 -1.555 -0.00007324 0.000008448 fixed
male -0.2616 0.02657 -9.846 -0.3137 -0.2095 fixed
sibling_count3 -0.04465 0.03971 -1.124 -0.1225 0.03318 fixed
sibling_count4 -0.08125 0.04277 -1.899 -0.1651 0.002591 fixed
sibling_count5 -0.115 0.05 -2.3 -0.213 -0.01698 fixed
sibling_count5+ -0.1278 0.04315 -2.963 -0.2124 -0.04327 fixed
sd_(Intercept).mother_pidlink 0.1898 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9926 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.06 0.4544 2.334 0.1698 1.951 fixed
birth_order 0.01785 0.008858 2.016 0.000492 0.03521 fixed
poly(age, 3, raw = TRUE)1 -0.09181 0.05156 -1.781 -0.1929 0.009241 fixed
poly(age, 3, raw = TRUE)2 0.003073 0.00184 1.671 -0.0005322 0.006679 fixed
poly(age, 3, raw = TRUE)3 -0.00003172 0.00002084 -1.522 -0.00007256 0.000009117 fixed
male -0.262 0.02656 -9.864 -0.3141 -0.2099 fixed
sibling_count3 -0.05323 0.03994 -1.333 -0.1315 0.02505 fixed
sibling_count4 -0.1011 0.0439 -2.303 -0.1871 -0.01505 fixed
sibling_count5 -0.1465 0.0524 -2.795 -0.2492 -0.04378 fixed
sibling_count5+ -0.1938 0.05419 -3.576 -0.3 -0.08757 fixed
sd_(Intercept).mother_pidlink 0.1929 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9917 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.044 0.4546 2.297 0.1532 1.935 fixed
poly(age, 3, raw = TRUE)1 -0.08873 0.05154 -1.722 -0.1898 0.01229 fixed
poly(age, 3, raw = TRUE)2 0.002964 0.001839 1.612 -0.0006406 0.006569 fixed
poly(age, 3, raw = TRUE)3 -0.00003055 0.00002083 -1.466 -0.00007138 0.00001029 fixed
male -0.2626 0.02655 -9.892 -0.3147 -0.2106 fixed
sibling_count3 -0.05612 0.04085 -1.374 -0.1362 0.02394 fixed
sibling_count4 -0.124 0.04582 -2.706 -0.2138 -0.03419 fixed
sibling_count5 -0.1948 0.05506 -3.537 -0.3027 -0.08684 fixed
sibling_count5+ -0.2001 0.05603 -3.572 -0.31 -0.09033 fixed
birth_order_nonlinear2 0.04114 0.03439 1.196 -0.02627 0.1086 fixed
birth_order_nonlinear3 0.05162 0.04243 1.217 -0.03154 0.1348 fixed
birth_order_nonlinear4 0.1592 0.05384 2.957 0.05366 0.2647 fixed
birth_order_nonlinear5 0.2166 0.06823 3.175 0.0829 0.3503 fixed
birth_order_nonlinear5+ 0.04953 0.06681 0.7413 -0.08142 0.1805 fixed
sd_(Intercept).mother_pidlink 0.1887 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9919 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.015 0.4558 2.227 0.1216 1.908 fixed
poly(age, 3, raw = TRUE)1 -0.08443 0.05168 -1.634 -0.1857 0.01686 fixed
poly(age, 3, raw = TRUE)2 0.002813 0.001844 1.525 -0.0008021 0.006427 fixed
poly(age, 3, raw = TRUE)3 -0.00002887 0.0000209 -1.382 -0.00006982 0.00001209 fixed
male -0.2616 0.02658 -9.842 -0.3137 -0.2095 fixed
count_birth_order2/2 0.01484 0.06064 0.2447 -0.104 0.1337 fixed
count_birth_order1/3 -0.09202 0.05348 -1.721 -0.1968 0.0128 fixed
count_birth_order2/3 0.009075 0.05919 0.1533 -0.1069 0.1251 fixed
count_birth_order3/3 -0.006927 0.06507 -0.1065 -0.1345 0.1206 fixed
count_birth_order1/4 -0.1042 0.06656 -1.565 -0.2346 0.02629 fixed
count_birth_order2/4 -0.1136 0.06851 -1.658 -0.2479 0.02067 fixed
count_birth_order3/4 -0.08828 0.07183 -1.229 -0.2291 0.0525 fixed
count_birth_order4/4 0.02013 0.07593 0.2651 -0.1287 0.169 fixed
count_birth_order1/5 -0.1585 0.08983 -1.765 -0.3346 0.01752 fixed
count_birth_order2/5 -0.151 0.09939 -1.519 -0.3458 0.04379 fixed
count_birth_order3/5 -0.1459 0.09485 -1.538 -0.3318 0.04 fixed
count_birth_order4/5 -0.05101 0.09137 -0.5583 -0.2301 0.1281 fixed
count_birth_order5/5 -0.05376 0.09711 -0.5536 -0.2441 0.1366 fixed
count_birth_order1/5+ -0.2499 0.09145 -2.732 -0.4291 -0.07063 fixed
count_birth_order2/5+ -0.1872 0.09072 -2.063 -0.365 -0.009383 fixed
count_birth_order3/5+ -0.1664 0.08997 -1.85 -0.3428 0.009894 fixed
count_birth_order4/5+ -0.03704 0.08803 -0.4208 -0.2096 0.1355 fixed
count_birth_order5/5+ 0.04899 0.0809 0.6055 -0.1096 0.2076 fixed
count_birth_order5+/5+ -0.1598 0.05974 -2.675 -0.2769 -0.04272 fixed
sd_(Intercept).mother_pidlink 0.1888 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9924 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 16605 16679 -8292 16583 NA NA NA
12 16603 16683 -8290 16579 4.058 1 0.04397
16 16600 16706 -8284 16568 11.43 4 0.0221
26 16617 16790 -8282 16565 3.206 10 0.9761

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Neuroticism

birthorder <- birthorder %>% mutate(outcome = big5_neu)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.601 0.1581 3.801 0.2911 0.911 fixed
poly(age, 3, raw = TRUE)1 -0.03614 0.01513 -2.388 -0.0658 -0.006479 fixed
poly(age, 3, raw = TRUE)2 0.0004516 0.0004443 1.016 -0.0004192 0.001322 fixed
poly(age, 3, raw = TRUE)3 -0.0000006353 0.000004084 -0.1556 -0.000008639 0.000007369 fixed
male 0.05519 0.01681 3.283 0.02224 0.08815 fixed
sibling_count3 0.07441 0.03286 2.265 0.01002 0.1388 fixed
sibling_count4 0.03297 0.03366 0.9794 -0.033 0.09893 fixed
sibling_count5 0.02561 0.03475 0.737 -0.04249 0.09372 fixed
sibling_count5+ 0.004546 0.02748 0.1654 -0.04932 0.05841 fixed
sd_(Intercept).mother_pidlink 0.1577 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9789 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.6006 0.1582 3.797 0.2906 0.9106 fixed
birth_order -0.0007634 0.003442 -0.2218 -0.00751 0.005983 fixed
poly(age, 3, raw = TRUE)1 -0.03592 0.01517 -2.368 -0.06565 -0.006189 fixed
poly(age, 3, raw = TRUE)2 0.0004439 0.0004457 0.996 -0.0004296 0.001317 fixed
poly(age, 3, raw = TRUE)3 -0.0000005662 0.000004096 -0.1382 -0.000008594 0.000007461 fixed
male 0.05521 0.01681 3.284 0.02226 0.08817 fixed
sibling_count3 0.07462 0.03287 2.27 0.01019 0.139 fixed
sibling_count4 0.03352 0.03375 0.9932 -0.03263 0.09968 fixed
sibling_count5 0.02656 0.03501 0.7586 -0.04206 0.09518 fixed
sibling_count5+ 0.007498 0.03054 0.2456 -0.05235 0.06735 fixed
sd_(Intercept).mother_pidlink 0.1578 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9789 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.6011 0.1586 3.79 0.2902 0.9119 fixed
poly(age, 3, raw = TRUE)1 -0.03675 0.01517 -2.422 -0.06649 -0.007014 fixed
poly(age, 3, raw = TRUE)2 0.0004618 0.0004458 1.036 -0.0004119 0.001336 fixed
poly(age, 3, raw = TRUE)3 -0.0000006348 0.000004097 -0.1549 -0.000008665 0.000007395 fixed
male 0.05494 0.01681 3.267 0.02198 0.08789 fixed
sibling_count3 0.06849 0.0334 2.05 0.003023 0.134 fixed
sibling_count4 0.02467 0.03487 0.7076 -0.04366 0.09301 fixed
sibling_count5 0.007275 0.0366 0.1987 -0.06446 0.07901 fixed
sibling_count5+ -0.007381 0.03233 -0.2283 -0.07074 0.05598 fixed
birth_order_nonlinear2 0.02011 0.02471 0.8138 -0.02832 0.06853 fixed
birth_order_nonlinear3 0.0327 0.02917 1.121 -0.02447 0.08988 fixed
birth_order_nonlinear4 0.02027 0.03321 0.6103 -0.04482 0.08535 fixed
birth_order_nonlinear5 0.07457 0.03775 1.975 0.000577 0.1486 fixed
birth_order_nonlinear5+ 0.008219 0.03081 0.2668 -0.05217 0.06861 fixed
sd_(Intercept).mother_pidlink 0.1577 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9789 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.5956 0.1593 3.739 0.2834 0.9079 fixed
poly(age, 3, raw = TRUE)1 -0.0371 0.01518 -2.444 -0.06685 -0.007352 fixed
poly(age, 3, raw = TRUE)2 0.0004657 0.0004459 1.044 -0.0004083 0.00134 fixed
poly(age, 3, raw = TRUE)3 -0.0000006162 0.000004098 -0.1504 -0.000008649 0.000007416 fixed
male 0.05477 0.01682 3.256 0.0218 0.08774 fixed
count_birth_order2/2 0.04968 0.04808 1.033 -0.04456 0.1439 fixed
count_birth_order1/3 0.09649 0.04475 2.156 0.008776 0.1842 fixed
count_birth_order2/3 0.05696 0.05006 1.138 -0.04115 0.1551 fixed
count_birth_order3/3 0.1377 0.05609 2.454 0.02774 0.2476 fixed
count_birth_order1/4 0.01046 0.05112 0.2045 -0.08975 0.1107 fixed
count_birth_order2/4 0.05887 0.05377 1.095 -0.04652 0.1643 fixed
count_birth_order3/4 0.08505 0.05844 1.455 -0.02949 0.1996 fixed
count_birth_order4/4 0.07642 0.06189 1.235 -0.04488 0.1977 fixed
count_birth_order1/5 0.004727 0.05803 0.08145 -0.109 0.1185 fixed
count_birth_order2/5 0.02951 0.06106 0.4834 -0.09016 0.1492 fixed
count_birth_order3/5 0.04063 0.06268 0.6482 -0.08222 0.1635 fixed
count_birth_order4/5 0.09274 0.06641 1.397 -0.03742 0.2229 fixed
count_birth_order5/5 0.08234 0.06789 1.213 -0.05072 0.2154 fixed
count_birth_order1/5+ 0.02683 0.04688 0.5724 -0.06505 0.1187 fixed
count_birth_order2/5+ 0.04649 0.04837 0.9612 -0.04831 0.1413 fixed
count_birth_order3/5+ 0.01851 0.0474 0.3905 -0.07439 0.1114 fixed
count_birth_order4/5+ -0.00249 0.04644 -0.05362 -0.09352 0.08854 fixed
count_birth_order5/5+ 0.08246 0.0468 1.762 -0.00926 0.1742 fixed
count_birth_order5+/5+ 0.01226 0.03604 0.34 -0.05839 0.0829 fixed
sd_(Intercept).mother_pidlink 0.1573 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9791 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 39311 39394 -19645 39289 NA NA NA
12 39313 39404 -19645 39289 0.04923 1 0.8244
16 39316 39437 -19642 39284 4.979 4 0.2895
26 39331 39527 -19639 39279 5.377 10 0.8646

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.07793 0.4239 0.1838 -0.7529 0.9088 fixed
poly(age, 3, raw = TRUE)1 0.02914 0.04807 0.6062 -0.06508 0.1234 fixed
poly(age, 3, raw = TRUE)2 -0.002039 0.001714 -1.19 -0.005399 0.001321 fixed
poly(age, 3, raw = TRUE)3 0.00002818 0.0000194 1.453 -0.000009837 0.00006621 fixed
male 0.03821 0.02483 1.539 -0.01044 0.08687 fixed
sibling_count3 0.09336 0.03775 2.473 0.01937 0.1673 fixed
sibling_count4 0.06312 0.04032 1.565 -0.01591 0.1421 fixed
sibling_count5 0.06854 0.04567 1.501 -0.02096 0.1581 fixed
sibling_count5+ 0.0178 0.03989 0.4463 -0.06038 0.09599 fixed
sd_(Intercept).mother_pidlink 0.1006 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9484 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.08652 0.4239 0.2041 -0.7444 0.9174 fixed
birth_order -0.01 0.008074 -1.238 -0.02583 0.005826 fixed
poly(age, 3, raw = TRUE)1 0.02948 0.04807 0.6132 -0.06474 0.1237 fixed
poly(age, 3, raw = TRUE)2 -0.002033 0.001714 -1.186 -0.005393 0.001327 fixed
poly(age, 3, raw = TRUE)3 0.00002773 0.0000194 1.429 -0.00001029 0.00006576 fixed
male 0.03862 0.02483 1.555 -0.01004 0.08728 fixed
sibling_count3 0.09815 0.03795 2.586 0.02377 0.1725 fixed
sibling_count4 0.07431 0.04132 1.798 -0.006685 0.1553 fixed
sibling_count5 0.08697 0.04803 1.811 -0.007179 0.1811 fixed
sibling_count5+ 0.05496 0.04992 1.101 -0.04289 0.1528 fixed
sd_(Intercept).mother_pidlink 0.1019 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9482 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.07131 0.4245 0.168 -0.7607 0.9034 fixed
poly(age, 3, raw = TRUE)1 0.0301 0.0481 0.6259 -0.06416 0.1244 fixed
poly(age, 3, raw = TRUE)2 -0.002058 0.001715 -1.2 -0.00542 0.001304 fixed
poly(age, 3, raw = TRUE)3 0.00002808 0.00001942 1.446 -0.000009974 0.00006614 fixed
male 0.03899 0.02483 1.57 -0.009682 0.08766 fixed
sibling_count3 0.09096 0.03885 2.341 0.0148 0.1671 fixed
sibling_count4 0.07859 0.04319 1.819 -0.006069 0.1633 fixed
sibling_count5 0.08039 0.05089 1.58 -0.01935 0.1801 fixed
sibling_count5+ 0.04368 0.05169 0.8451 -0.05762 0.145 fixed
birth_order_nonlinear2 -0.01106 0.03269 -0.3382 -0.07513 0.05301 fixed
birth_order_nonlinear3 0.01072 0.04018 0.2668 -0.06803 0.08947 fixed
birth_order_nonlinear4 -0.08731 0.04954 -1.762 -0.1844 0.00979 fixed
birth_order_nonlinear5 0.02271 0.06194 0.3667 -0.0987 0.1441 fixed
birth_order_nonlinear5+ -0.0531 0.06103 -0.8702 -0.1727 0.06651 fixed
sd_(Intercept).mother_pidlink 0.1078 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9476 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.07141 0.4255 0.1678 -0.7625 0.9053 fixed
poly(age, 3, raw = TRUE)1 0.02803 0.04821 0.5815 -0.06646 0.1225 fixed
poly(age, 3, raw = TRUE)2 -0.001994 0.00172 -1.159 -0.005364 0.001377 fixed
poly(age, 3, raw = TRUE)3 0.00002749 0.00001947 1.412 -0.00001067 0.00006564 fixed
male 0.03836 0.02486 1.543 -0.01036 0.08708 fixed
count_birth_order2/2 0.04808 0.059 0.815 -0.06756 0.1637 fixed
count_birth_order1/3 0.1422 0.05114 2.781 0.04197 0.2424 fixed
count_birth_order2/3 0.08164 0.05592 1.46 -0.02797 0.1912 fixed
count_birth_order3/3 0.08769 0.06257 1.401 -0.03495 0.2103 fixed
count_birth_order1/4 0.1145 0.06254 1.831 -0.00807 0.2371 fixed
count_birth_order2/4 0.08613 0.06479 1.329 -0.04086 0.2131 fixed
count_birth_order3/4 0.07842 0.06855 1.144 -0.05593 0.2128 fixed
count_birth_order4/4 0.02476 0.07127 0.3474 -0.1149 0.1645 fixed
count_birth_order1/5 0.06937 0.08504 0.8157 -0.09731 0.236 fixed
count_birth_order2/5 0.03969 0.09114 0.4355 -0.1389 0.2183 fixed
count_birth_order3/5 0.1673 0.08546 1.958 -0.0002078 0.3348 fixed
count_birth_order4/5 0.0241 0.08231 0.2927 -0.1372 0.1854 fixed
count_birth_order5/5 0.1299 0.08569 1.516 -0.03808 0.2978 fixed
count_birth_order1/5+ 0.0308 0.08431 0.3653 -0.1344 0.196 fixed
count_birth_order2/5+ 0.05316 0.08353 0.6365 -0.1105 0.2169 fixed
count_birth_order3/5+ 0.1418 0.08377 1.693 -0.02237 0.306 fixed
count_birth_order4/5+ -0.0495 0.07896 -0.6269 -0.2043 0.1053 fixed
count_birth_order5/5+ 0.08194 0.07463 1.098 -0.06433 0.2282 fixed
count_birth_order5+/5+ 0.01088 0.05522 0.1971 -0.09735 0.1191 fixed
sd_(Intercept).mother_pidlink 0.1066 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9481 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 16263 16336 -8120 16241 NA NA NA
12 16263 16343 -8120 16239 1.535 1 0.2154
16 16268 16375 -8118 16236 3.665 4 0.4532
26 16283 16456 -8115 16231 4.968 10 0.8933

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.0317 0.4229 0.07496 -0.7972 0.8606 fixed
poly(age, 3, raw = TRUE)1 0.03346 0.04798 0.6974 -0.06058 0.1275 fixed
poly(age, 3, raw = TRUE)2 -0.002179 0.001712 -1.273 -0.005534 0.001175 fixed
poly(age, 3, raw = TRUE)3 0.00002954 0.00001937 1.525 -0.000008427 0.00006751 fixed
male 0.04097 0.02475 1.655 -0.007543 0.08948 fixed
sibling_count3 0.08282 0.04092 2.024 0.002613 0.163 fixed
sibling_count4 0.06103 0.04287 1.424 -0.02299 0.145 fixed
sibling_count5 0.0811 0.04553 1.781 -0.008131 0.1703 fixed
sibling_count5+ 0.03651 0.03994 0.9141 -0.04178 0.1148 fixed
sd_(Intercept).mother_pidlink 0.1005 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9493 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.04441 0.4229 0.105 -0.7845 0.8733 fixed
birth_order -0.01109 0.00701 -1.582 -0.02483 0.00265 fixed
poly(age, 3, raw = TRUE)1 0.03346 0.04798 0.6974 -0.06057 0.1275 fixed
poly(age, 3, raw = TRUE)2 -0.002159 0.001711 -1.261 -0.005513 0.001196 fixed
poly(age, 3, raw = TRUE)3 0.00002889 0.00001937 1.491 -0.000009077 0.00006686 fixed
male 0.0413 0.02475 1.669 -0.007202 0.08981 fixed
sibling_count3 0.0881 0.04106 2.146 0.007622 0.1686 fixed
sibling_count4 0.07293 0.04353 1.675 -0.01238 0.1582 fixed
sibling_count5 0.09992 0.04706 2.123 0.007676 0.1922 fixed
sibling_count5+ 0.07683 0.0474 1.621 -0.01606 0.1697 fixed
sd_(Intercept).mother_pidlink 0.1033 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9489 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.03876 0.4231 0.09161 -0.7906 0.8681 fixed
poly(age, 3, raw = TRUE)1 0.03369 0.04797 0.7023 -0.06032 0.1277 fixed
poly(age, 3, raw = TRUE)2 -0.002167 0.001711 -1.266 -0.005521 0.001187 fixed
poly(age, 3, raw = TRUE)3 0.000029 0.00001937 1.497 -0.000008971 0.00006698 fixed
male 0.04192 0.02474 1.695 -0.006566 0.0904 fixed
sibling_count3 0.07471 0.04189 1.784 -0.007392 0.1568 fixed
sibling_count4 0.0765 0.04525 1.69 -0.01219 0.1652 fixed
sibling_count5 0.08472 0.04963 1.707 -0.01255 0.182 fixed
sibling_count5+ 0.06516 0.0491 1.327 -0.03107 0.1614 fixed
birth_order_nonlinear2 -0.03395 0.03325 -1.021 -0.09911 0.03121 fixed
birth_order_nonlinear3 0.0351 0.04 0.8776 -0.04329 0.1135 fixed
birth_order_nonlinear4 -0.1214 0.04801 -2.528 -0.2155 -0.02729 fixed
birth_order_nonlinear5 0.04432 0.0589 0.7524 -0.07113 0.1598 fixed
birth_order_nonlinear5+ -0.07547 0.05434 -1.389 -0.182 0.03104 fixed
sd_(Intercept).mother_pidlink 0.1105 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9475 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.03366 0.424 0.07939 -0.7973 0.8646 fixed
poly(age, 3, raw = TRUE)1 0.03155 0.04804 0.6567 -0.06261 0.1257 fixed
poly(age, 3, raw = TRUE)2 -0.002097 0.001714 -1.223 -0.005457 0.001263 fixed
poly(age, 3, raw = TRUE)3 0.00002826 0.00001941 1.456 -0.000009788 0.00006631 fixed
male 0.04179 0.02476 1.688 -0.00673 0.09032 fixed
count_birth_order2/2 0.03961 0.0645 0.6141 -0.0868 0.166 fixed
count_birth_order1/3 0.1374 0.05538 2.481 0.02883 0.2459 fixed
count_birth_order2/3 0.0286 0.0601 0.4759 -0.0892 0.1464 fixed
count_birth_order3/3 0.1168 0.0677 1.725 -0.01592 0.2495 fixed
count_birth_order1/4 0.0847 0.06551 1.293 -0.04369 0.2131 fixed
count_birth_order2/4 0.05894 0.0668 0.8824 -0.07198 0.1899 fixed
count_birth_order3/4 0.1447 0.0734 1.971 0.0008398 0.2886 fixed
count_birth_order4/4 0.01153 0.07568 0.1524 -0.1368 0.1599 fixed
count_birth_order1/5 0.09265 0.07779 1.191 -0.05981 0.2451 fixed
count_birth_order2/5 0.09271 0.08364 1.109 -0.07121 0.2566 fixed
count_birth_order3/5 0.1264 0.08136 1.553 -0.03309 0.2859 fixed
count_birth_order4/5 0.04972 0.08439 0.5892 -0.1157 0.2151 fixed
count_birth_order5/5 0.1191 0.08463 1.407 -0.04679 0.2849 fixed
count_birth_order1/5+ 0.1082 0.0739 1.465 -0.03659 0.2531 fixed
count_birth_order2/5+ 0.0554 0.07728 0.7169 -0.09607 0.2069 fixed
count_birth_order3/5+ 0.1571 0.07588 2.07 0.008331 0.3058 fixed
count_birth_order4/5+ -0.09833 0.07306 -1.346 -0.2415 0.04487 fixed
count_birth_order5/5+ 0.1605 0.07513 2.136 0.01323 0.3078 fixed
count_birth_order5+/5+ 0.01527 0.05421 0.2817 -0.09098 0.1215 fixed
sd_(Intercept).mother_pidlink 0.1101 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9479 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 16414 16488 -8196 16392 NA NA NA
12 16414 16494 -8195 16390 2.502 1 0.1137
16 16410 16517 -8189 16378 11.42 4 0.02226
26 16424 16598 -8186 16372 6.095 10 0.8072

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.2053 0.4296 0.4778 -0.6367 1.047 fixed
poly(age, 3, raw = TRUE)1 0.01525 0.04874 0.3129 -0.08027 0.1108 fixed
poly(age, 3, raw = TRUE)2 -0.00156 0.001739 -0.8973 -0.004968 0.001848 fixed
poly(age, 3, raw = TRUE)3 0.0000229 0.00001968 1.163 -0.00001569 0.00006148 fixed
male 0.0369 0.02513 1.468 -0.01235 0.08614 fixed
sibling_count3 0.08947 0.03732 2.398 0.01633 0.1626 fixed
sibling_count4 0.03919 0.04012 0.9767 -0.03945 0.1178 fixed
sibling_count5 0.05698 0.04679 1.218 -0.03473 0.1487 fixed
sibling_count5+ 0.03513 0.04032 0.8712 -0.0439 0.1142 fixed
sd_(Intercept).mother_pidlink 0.1033 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9496 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.214 0.4296 0.4982 -0.6279 1.056 fixed
birth_order -0.01305 0.008345 -1.564 -0.02941 0.003305 fixed
poly(age, 3, raw = TRUE)1 0.01595 0.04873 0.3273 -0.07957 0.1115 fixed
poly(age, 3, raw = TRUE)2 -0.00156 0.001738 -0.8971 -0.004967 0.001848 fixed
poly(age, 3, raw = TRUE)3 0.00002239 0.00001969 1.137 -0.00001619 0.00006097 fixed
male 0.0372 0.02513 1.481 -0.01205 0.08644 fixed
sibling_count3 0.09572 0.03753 2.551 0.02217 0.1693 fixed
sibling_count4 0.05364 0.04117 1.303 -0.02705 0.1343 fixed
sibling_count5 0.07992 0.04903 1.63 -0.01618 0.176 fixed
sibling_count5+ 0.0833 0.05074 1.642 -0.01615 0.1828 fixed
sd_(Intercept).mother_pidlink 0.1035 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9495 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.1918 0.4301 0.446 -0.6511 1.035 fixed
poly(age, 3, raw = TRUE)1 0.01694 0.04875 0.3474 -0.07861 0.1125 fixed
poly(age, 3, raw = TRUE)2 -0.001603 0.001739 -0.9217 -0.005012 0.001806 fixed
poly(age, 3, raw = TRUE)3 0.00002304 0.0000197 1.17 -0.00001556 0.00006165 fixed
male 0.03758 0.02513 1.495 -0.01167 0.08683 fixed
sibling_count3 0.08919 0.03845 2.319 0.01382 0.1646 fixed
sibling_count4 0.05798 0.04308 1.346 -0.02646 0.1424 fixed
sibling_count5 0.06803 0.05169 1.316 -0.03328 0.1693 fixed
sibling_count5+ 0.06393 0.05261 1.215 -0.03918 0.167 fixed
birth_order_nonlinear2 -0.009548 0.03268 -0.2922 -0.0736 0.05451 fixed
birth_order_nonlinear3 0.001187 0.04029 0.02947 -0.07777 0.08015 fixed
birth_order_nonlinear4 -0.09376 0.05109 -1.835 -0.1939 0.006373 fixed
birth_order_nonlinear5 0.04599 0.06475 0.7104 -0.08091 0.1729 fixed
birth_order_nonlinear5+ -0.06619 0.06315 -1.048 -0.19 0.05759 fixed
sd_(Intercept).mother_pidlink 0.1082 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.949 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.1918 0.4311 0.4448 -0.6532 1.037 fixed
poly(age, 3, raw = TRUE)1 0.01537 0.04887 0.3146 -0.08041 0.1112 fixed
poly(age, 3, raw = TRUE)2 -0.001561 0.001744 -0.8951 -0.004978 0.001857 fixed
poly(age, 3, raw = TRUE)3 0.00002274 0.00001975 1.151 -0.00001597 0.00006145 fixed
male 0.03747 0.02516 1.49 -0.01183 0.08678 fixed
count_birth_order2/2 0.03766 0.0576 0.6538 -0.07524 0.1505 fixed
count_birth_order1/3 0.1411 0.05059 2.789 0.04195 0.2402 fixed
count_birth_order2/3 0.07607 0.056 1.358 -0.03368 0.1858 fixed
count_birth_order3/3 0.06606 0.06157 1.073 -0.05462 0.1867 fixed
count_birth_order1/4 0.08699 0.06296 1.382 -0.03641 0.2104 fixed
count_birth_order2/4 0.05859 0.06481 0.904 -0.06844 0.1856 fixed
count_birth_order3/4 0.06053 0.06797 0.8905 -0.07269 0.1937 fixed
count_birth_order4/4 -0.01195 0.07185 -0.1663 -0.1528 0.1289 fixed
count_birth_order1/5 0.06638 0.08497 0.7812 -0.1002 0.2329 fixed
count_birth_order2/5 0.05185 0.09403 0.5514 -0.1324 0.2362 fixed
count_birth_order3/5 0.1164 0.08975 1.297 -0.05947 0.2923 fixed
count_birth_order4/5 0.004368 0.08646 0.05052 -0.1651 0.1738 fixed
count_birth_order5/5 0.1259 0.09189 1.37 -0.05423 0.306 fixed
count_birth_order1/5+ 0.002119 0.08648 0.02451 -0.1674 0.1716 fixed
count_birth_order2/5+ 0.07657 0.08579 0.8925 -0.09158 0.2447 fixed
count_birth_order3/5+ 0.1699 0.08512 1.996 0.003086 0.3367 fixed
count_birth_order4/5+ -0.03566 0.08329 -0.4281 -0.1989 0.1276 fixed
count_birth_order5/5+ 0.1293 0.07655 1.689 -0.02071 0.2794 fixed
count_birth_order5+/5+ 0.01442 0.05626 0.2563 -0.09585 0.1247 fixed
sd_(Intercept).mother_pidlink 0.106 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9496 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 15955 16028 -7966 15933 NA NA NA
12 15954 16034 -7965 15930 2.449 1 0.1176
16 15958 16065 -7963 15926 3.912 4 0.418
26 15973 16146 -7961 15921 5.381 10 0.8643

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Conscientiousness

birthorder <- birthorder %>% mutate(outcome = big5_con)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.384 0.1564 -8.846 -1.69 -1.077 fixed
poly(age, 3, raw = TRUE)1 0.07934 0.01497 5.299 0.04999 0.1087 fixed
poly(age, 3, raw = TRUE)2 -0.001195 0.0004398 -2.718 -0.002057 -0.0003333 fixed
poly(age, 3, raw = TRUE)3 0.000005148 0.000004043 1.273 -0.000002777 0.00001307 fixed
male 0.0443 0.01659 2.669 0.01177 0.07682 fixed
sibling_count3 -0.01129 0.03263 -0.346 -0.07525 0.05267 fixed
sibling_count4 -0.01053 0.03347 -0.3145 -0.07613 0.05508 fixed
sibling_count5 -0.0000886 0.03461 -0.00256 -0.06792 0.06775 fixed
sibling_count5+ 0.02139 0.02733 0.7829 -0.03217 0.07495 fixed
sd_(Intercept).mother_pidlink 0.2052 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9578 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.384 0.1564 -8.85 -1.691 -1.078 fixed
birth_order -0.001557 0.003422 -0.4549 -0.008263 0.00515 fixed
poly(age, 3, raw = TRUE)1 0.0798 0.01501 5.317 0.05038 0.1092 fixed
poly(age, 3, raw = TRUE)2 -0.001211 0.0004412 -2.745 -0.002076 -0.0003464 fixed
poly(age, 3, raw = TRUE)3 0.000005292 0.000004056 1.305 -0.000002658 0.00001324 fixed
male 0.04434 0.0166 2.672 0.01181 0.07687 fixed
sibling_count3 -0.01089 0.03265 -0.3337 -0.07488 0.05309 fixed
sibling_count4 -0.009402 0.03357 -0.2801 -0.07519 0.05639 fixed
sibling_count5 0.00183 0.03487 0.05247 -0.06652 0.07017 fixed
sibling_count5+ 0.02738 0.03034 0.9025 -0.03208 0.08683 fixed
sd_(Intercept).mother_pidlink 0.2055 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9578 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.373 0.1568 -8.753 -1.68 -1.066 fixed
poly(age, 3, raw = TRUE)1 0.07988 0.01501 5.32 0.05045 0.1093 fixed
poly(age, 3, raw = TRUE)2 -0.001203 0.0004413 -2.726 -0.002068 -0.0003381 fixed
poly(age, 3, raw = TRUE)3 0.000005103 0.000004057 1.258 -0.000002848 0.00001305 fixed
male 0.04452 0.0166 2.682 0.01199 0.07704 fixed
sibling_count3 -0.004833 0.03316 -0.1457 -0.06983 0.06016 fixed
sibling_count4 -0.003616 0.03464 -0.1044 -0.07151 0.06427 fixed
sibling_count5 0.01333 0.0364 0.3662 -0.05802 0.08468 fixed
sibling_count5+ 0.0351 0.03206 1.095 -0.02774 0.09794 fixed
birth_order_nonlinear2 -0.04456 0.02432 -1.832 -0.09222 0.003106 fixed
birth_order_nonlinear3 -0.04247 0.02871 -1.479 -0.09874 0.01381 fixed
birth_order_nonlinear4 -0.01607 0.03269 -0.4915 -0.08014 0.048 fixed
birth_order_nonlinear5 -0.06266 0.03717 -1.686 -0.1355 0.01018 fixed
birth_order_nonlinear5+ -0.03018 0.03045 -0.9914 -0.08986 0.02949 fixed
sd_(Intercept).mother_pidlink 0.2051 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9578 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.345 0.1575 -8.541 -1.654 -1.037 fixed
poly(age, 3, raw = TRUE)1 0.08008 0.01501 5.333 0.05065 0.1095 fixed
poly(age, 3, raw = TRUE)2 -0.001199 0.0004413 -2.717 -0.002064 -0.0003343 fixed
poly(age, 3, raw = TRUE)3 0.000004988 0.000004057 1.229 -0.000002965 0.00001294 fixed
male 0.0447 0.0166 2.693 0.01217 0.07723 fixed
count_birth_order2/2 -0.1323 0.04729 -2.798 -0.225 -0.03964 fixed
count_birth_order1/3 -0.05977 0.0442 -1.352 -0.1464 0.02685 fixed
count_birth_order2/3 -0.05331 0.04943 -1.078 -0.1502 0.04357 fixed
count_birth_order3/3 -0.07588 0.05538 -1.37 -0.1844 0.03267 fixed
count_birth_order1/4 -0.03678 0.05048 -0.7286 -0.1357 0.06216 fixed
count_birth_order2/4 -0.08135 0.0531 -1.532 -0.1854 0.02271 fixed
count_birth_order3/4 -0.0356 0.0577 -0.6169 -0.1487 0.07749 fixed
count_birth_order4/4 -0.1034 0.0611 -1.692 -0.2231 0.01639 fixed
count_birth_order1/5 0.02975 0.0573 0.5192 -0.08256 0.1421 fixed
count_birth_order2/5 -0.03593 0.06028 -0.596 -0.1541 0.08222 fixed
count_birth_order3/5 -0.08495 0.06188 -1.373 -0.2062 0.03633 fixed
count_birth_order4/5 -0.04652 0.06556 -0.7095 -0.175 0.08198 fixed
count_birth_order5/5 -0.1564 0.06702 -2.333 -0.2877 -0.02501 fixed
count_birth_order1/5+ -0.04653 0.04629 -1.005 -0.1372 0.04419 fixed
count_birth_order2/5+ -0.02981 0.04775 -0.6243 -0.1234 0.06378 fixed
count_birth_order3/5+ -0.05692 0.04679 -1.217 -0.1486 0.03478 fixed
count_birth_order4/5+ 0.01003 0.04585 0.2187 -0.07983 0.09989 fixed
count_birth_order5/5+ -0.03622 0.04619 -0.7841 -0.1268 0.05432 fixed
count_birth_order5+/5+ -0.02832 0.0357 -0.7933 -0.0983 0.04165 fixed
sd_(Intercept).mother_pidlink 0.2051 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9577 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 38959 39042 -19468 38937 NA NA NA
12 38960 39051 -19468 38936 0.2064 1 0.6496
16 38963 39084 -19466 38931 5.284 4 0.2594
26 38971 39167 -19459 38919 12.56 10 0.2493

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.8647 0.4405 -1.963 -1.728 -0.001376 fixed
poly(age, 3, raw = TRUE)1 0.01416 0.04997 0.2834 -0.08377 0.1121 fixed
poly(age, 3, raw = TRUE)2 0.001255 0.001782 0.7039 -0.002239 0.004748 fixed
poly(age, 3, raw = TRUE)3 -0.00002195 0.00002018 -1.088 -0.0000615 0.0000176 fixed
male 0.05224 0.02578 2.026 0.001712 0.1028 fixed
sibling_count3 -0.02611 0.03958 -0.6597 -0.1037 0.05147 fixed
sibling_count4 -0.04355 0.0424 -1.027 -0.1266 0.03955 fixed
sibling_count5 -0.01096 0.04815 -0.2276 -0.1053 0.08342 fixed
sibling_count5+ 0.03937 0.04215 0.9341 -0.04324 0.122 fixed
sd_(Intercept).mother_pidlink 0.2157 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9676 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.85 0.4404 -1.93 -1.713 0.01315 fixed
birth_order -0.01736 0.008434 -2.058 -0.03389 -0.0008289 fixed
poly(age, 3, raw = TRUE)1 0.01482 0.04995 0.2967 -0.08309 0.1127 fixed
poly(age, 3, raw = TRUE)2 0.001262 0.001782 0.7084 -0.00223 0.004755 fixed
poly(age, 3, raw = TRUE)3 -0.0000227 0.00002018 -1.125 -0.00006225 0.00001684 fixed
male 0.05295 0.02577 2.054 0.002433 0.1035 fixed
sibling_count3 -0.01776 0.03978 -0.4464 -0.09573 0.06021 fixed
sibling_count4 -0.02394 0.04345 -0.551 -0.1091 0.06122 fixed
sibling_count5 0.02124 0.05063 0.4196 -0.07798 0.1205 fixed
sibling_count5+ 0.1041 0.0526 1.98 0.00106 0.2072 fixed
sd_(Intercept).mother_pidlink 0.2163 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9672 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.8407 0.4411 -1.906 -1.705 0.02385 fixed
poly(age, 3, raw = TRUE)1 0.01289 0.04999 0.2579 -0.08509 0.1109 fixed
poly(age, 3, raw = TRUE)2 0.001337 0.001783 0.7497 -0.002159 0.004833 fixed
poly(age, 3, raw = TRUE)3 -0.00002361 0.00002019 -1.169 -0.00006319 0.00001597 fixed
male 0.05276 0.02578 2.046 0.002225 0.1033 fixed
sibling_count3 -0.01599 0.04069 -0.393 -0.09574 0.06376 fixed
sibling_count4 -0.01708 0.04533 -0.3768 -0.1059 0.07176 fixed
sibling_count5 0.02747 0.05349 0.5135 -0.07737 0.1323 fixed
sibling_count5+ 0.1085 0.05435 1.997 0.002011 0.215 fixed
birth_order_nonlinear2 -0.05026 0.03374 -1.49 -0.1164 0.01586 fixed
birth_order_nonlinear3 -0.04458 0.04152 -1.074 -0.126 0.0368 fixed
birth_order_nonlinear4 -0.08636 0.05124 -1.685 -0.1868 0.01408 fixed
birth_order_nonlinear5 -0.08063 0.06405 -1.259 -0.2062 0.04491 fixed
birth_order_nonlinear5+ -0.127 0.06351 -2 -0.2515 -0.002564 fixed
sd_(Intercept).mother_pidlink 0.2164 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9674 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.8614 0.4418 -1.95 -1.727 0.004539 fixed
poly(age, 3, raw = TRUE)1 0.01798 0.05007 0.3591 -0.08016 0.1161 fixed
poly(age, 3, raw = TRUE)2 0.001121 0.001787 0.6275 -0.002381 0.004623 fixed
poly(age, 3, raw = TRUE)3 -0.00002076 0.00002023 -1.026 -0.00006042 0.0000189 fixed
male 0.05421 0.02579 2.102 0.003657 0.1048 fixed
count_birth_order2/2 -0.1018 0.06088 -1.672 -0.2211 0.01753 fixed
count_birth_order1/3 -0.004458 0.05311 -0.08395 -0.1085 0.09963 fixed
count_birth_order2/3 -0.1003 0.05806 -1.727 -0.2141 0.01355 fixed
count_birth_order3/3 -0.1107 0.06495 -1.705 -0.238 0.01658 fixed
count_birth_order1/4 -0.05653 0.06495 -0.8703 -0.1838 0.07077 fixed
count_birth_order2/4 -0.05995 0.06727 -0.8911 -0.1918 0.0719 fixed
count_birth_order3/4 -0.02575 0.07114 -0.362 -0.1652 0.1137 fixed
count_birth_order4/4 -0.1802 0.07397 -2.437 -0.3252 -0.03527 fixed
count_birth_order1/5 -0.09011 0.0883 -1.021 -0.2632 0.08294 fixed
count_birth_order2/5 0.09637 0.0946 1.019 -0.08904 0.2818 fixed
count_birth_order3/5 0.02785 0.08869 0.314 -0.146 0.2017 fixed
count_birth_order4/5 -0.06482 0.08542 -0.7588 -0.2322 0.1026 fixed
count_birth_order5/5 -0.1599 0.08892 -1.798 -0.3341 0.01441 fixed
count_birth_order1/5+ 0.03187 0.08759 0.3638 -0.1398 0.2035 fixed
count_birth_order2/5+ 0.01605 0.08675 0.1851 -0.154 0.1861 fixed
count_birth_order3/5+ -0.03007 0.08697 -0.3458 -0.2005 0.1404 fixed
count_birth_order4/5+ 0.06919 0.08196 0.8442 -0.09145 0.2298 fixed
count_birth_order5/5+ 0.07304 0.07745 0.943 -0.07876 0.2249 fixed
count_birth_order5+/5+ -0.03498 0.05771 -0.6062 -0.1481 0.07812 fixed
sd_(Intercept).mother_pidlink 0.2171 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.967 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 16715 16788 -8346 16693 NA NA NA
12 16713 16793 -8344 16689 4.241 1 0.03947
16 16719 16826 -8344 16687 1.514 4 0.8241
26 16726 16900 -8337 16674 13.11 10 0.2177

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.8294 0.439 -1.889 -1.69 0.03101 fixed
poly(age, 3, raw = TRUE)1 0.01291 0.04982 0.259 -0.08475 0.1106 fixed
poly(age, 3, raw = TRUE)2 0.001314 0.001778 0.7389 -0.002171 0.004798 fixed
poly(age, 3, raw = TRUE)3 -0.00002253 0.00002013 -1.12 -0.00006198 0.00001692 fixed
male 0.05108 0.02567 1.99 0.0007665 0.1014 fixed
sibling_count3 -0.06959 0.04285 -1.624 -0.1536 0.01439 fixed
sibling_count4 -0.06697 0.04498 -1.489 -0.1551 0.0212 fixed
sibling_count5 -0.0582 0.04789 -1.215 -0.1521 0.03566 fixed
sibling_count5+ -0.01844 0.04201 -0.4389 -0.1008 0.06391 fixed
sd_(Intercept).mother_pidlink 0.2172 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9672 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.8205 0.4391 -1.869 -1.681 0.04006 fixed
birth_order -0.007815 0.007334 -1.066 -0.02219 0.00656 fixed
poly(age, 3, raw = TRUE)1 0.01293 0.04982 0.2596 -0.08472 0.1106 fixed
poly(age, 3, raw = TRUE)2 0.001327 0.001778 0.7463 -0.002158 0.004811 fixed
poly(age, 3, raw = TRUE)3 -0.00002298 0.00002013 -1.141 -0.00006244 0.00001648 fixed
male 0.05132 0.02567 1.999 0.001008 0.1016 fixed
sibling_count3 -0.06585 0.04299 -1.532 -0.1501 0.01842 fixed
sibling_count4 -0.0585 0.04568 -1.281 -0.148 0.03103 fixed
sibling_count5 -0.04485 0.04951 -0.9058 -0.1419 0.05219 fixed
sibling_count5+ 0.01003 0.0498 0.2014 -0.08757 0.1076 fixed
sd_(Intercept).mother_pidlink 0.2179 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.967 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.7972 0.4396 -1.813 -1.659 0.06439 fixed
poly(age, 3, raw = TRUE)1 0.01086 0.04984 0.2178 -0.08683 0.1085 fixed
poly(age, 3, raw = TRUE)2 0.001404 0.001779 0.7896 -0.002082 0.004891 fixed
poly(age, 3, raw = TRUE)3 -0.00002385 0.00002015 -1.184 -0.00006333 0.00001563 fixed
male 0.05093 0.02568 1.983 0.0006034 0.1013 fixed
sibling_count3 -0.05132 0.04384 -1.171 -0.1372 0.0346 fixed
sibling_count4 -0.04716 0.04742 -0.9943 -0.1401 0.04579 fixed
sibling_count5 -0.03619 0.0521 -0.6946 -0.1383 0.06592 fixed
sibling_count5+ 0.01453 0.0515 0.2822 -0.0864 0.1155 fixed
birth_order_nonlinear2 -0.0495 0.03432 -1.442 -0.1168 0.01776 fixed
birth_order_nonlinear3 -0.08248 0.04134 -1.995 -0.1635 -0.00145 fixed
birth_order_nonlinear4 -0.02223 0.04967 -0.4476 -0.1196 0.07511 fixed
birth_order_nonlinear5 -0.04183 0.06094 -0.6863 -0.1613 0.07762 fixed
birth_order_nonlinear5+ -0.0672 0.05659 -1.187 -0.1781 0.04372 fixed
sd_(Intercept).mother_pidlink 0.2171 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9672 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.8331 0.4403 -1.892 -1.696 0.02997 fixed
poly(age, 3, raw = TRUE)1 0.01647 0.04991 0.3301 -0.08135 0.1143 fixed
poly(age, 3, raw = TRUE)2 0.001182 0.001781 0.6636 -0.002309 0.004674 fixed
poly(age, 3, raw = TRUE)3 -0.00002112 0.00002018 -1.047 -0.00006067 0.00001843 fixed
male 0.0522 0.02569 2.032 0.001847 0.1026 fixed
count_birth_order2/2 -0.07531 0.06659 -1.131 -0.2058 0.05521 fixed
count_birth_order1/3 -0.04684 0.05752 -0.8144 -0.1596 0.06589 fixed
count_birth_order2/3 -0.1051 0.06241 -1.683 -0.2274 0.01726 fixed
count_birth_order3/3 -0.1749 0.07028 -2.489 -0.3127 -0.03718 fixed
count_birth_order1/4 -0.04383 0.06804 -0.6442 -0.1772 0.08952 fixed
count_birth_order2/4 -0.09198 0.06936 -1.326 -0.2279 0.04397 fixed
count_birth_order3/4 -0.08984 0.07619 -1.179 -0.2392 0.0595 fixed
count_birth_order4/4 -0.1675 0.07856 -2.132 -0.3215 -0.01353 fixed
count_birth_order1/5 -0.03052 0.08079 -0.3778 -0.1889 0.1278 fixed
count_birth_order2/5 -0.122 0.08684 -1.404 -0.2922 0.04824 fixed
count_birth_order3/5 -0.09178 0.08447 -1.087 -0.2573 0.07377 fixed
count_birth_order4/5 0.002615 0.08758 0.02985 -0.169 0.1743 fixed
count_birth_order5/5 -0.1843 0.08783 -2.098 -0.3564 -0.01215 fixed
count_birth_order1/5+ -0.07646 0.07677 -0.9959 -0.2269 0.07402 fixed
count_birth_order2/5+ -0.0221 0.08027 -0.2753 -0.1794 0.1352 fixed
count_birth_order3/5+ -0.1136 0.07879 -1.442 -0.268 0.04081 fixed
count_birth_order4/5+ 0.01616 0.07584 0.2131 -0.1325 0.1648 fixed
count_birth_order5/5+ 0.03516 0.07798 0.4509 -0.1177 0.188 fixed
count_birth_order5+/5+ -0.06057 0.05661 -1.07 -0.1715 0.05038 fixed
sd_(Intercept).mother_pidlink 0.2173 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9672 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 16858 16931 -8418 16836 NA NA NA
12 16859 16939 -8417 16835 1.136 1 0.2865
16 16863 16970 -8415 16831 3.922 4 0.4166
26 16873 17047 -8411 16821 9.357 10 0.4986

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.7476 0.4452 -1.679 -1.62 0.1249 fixed
poly(age, 3, raw = TRUE)1 0.003025 0.05052 0.05987 -0.096 0.102 fixed
poly(age, 3, raw = TRUE)2 0.001617 0.001803 0.8969 -0.001917 0.005151 fixed
poly(age, 3, raw = TRUE)3 -0.00002573 0.00002042 -1.26 -0.00006575 0.00001429 fixed
male 0.05245 0.02602 2.016 0.001454 0.1035 fixed
sibling_count3 -0.06345 0.03902 -1.626 -0.1399 0.01303 fixed
sibling_count4 -0.05627 0.04207 -1.338 -0.1387 0.02618 fixed
sibling_count5 0.01114 0.04923 0.2263 -0.08535 0.1076 fixed
sibling_count5+ 0.04261 0.04251 1.002 -0.0407 0.1259 fixed
sd_(Intercept).mother_pidlink 0.215 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9666 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.7362 0.4451 -1.654 -1.609 0.1362 fixed
birth_order -0.0169 0.008689 -1.945 -0.03393 0.0001265 fixed
poly(age, 3, raw = TRUE)1 0.003965 0.05051 0.07849 -0.09504 0.103 fixed
poly(age, 3, raw = TRUE)2 0.001616 0.001802 0.8963 -0.001917 0.005148 fixed
poly(age, 3, raw = TRUE)3 -0.00002637 0.00002042 -1.292 -0.00006639 0.00001365 fixed
male 0.05286 0.02601 2.032 0.001874 0.1038 fixed
sibling_count3 -0.0553 0.03924 -1.409 -0.1322 0.0216 fixed
sibling_count4 -0.03742 0.04316 -0.8669 -0.122 0.04718 fixed
sibling_count5 0.04108 0.05157 0.7966 -0.05999 0.1422 fixed
sibling_count5+ 0.1052 0.05331 1.974 0.0007359 0.2097 fixed
sd_(Intercept).mother_pidlink 0.2149 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9664 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.7235 0.4458 -1.623 -1.597 0.1503 fixed
poly(age, 3, raw = TRUE)1 0.001757 0.05055 0.03477 -0.09731 0.1008 fixed
poly(age, 3, raw = TRUE)2 0.001695 0.001804 0.9395 -0.001841 0.00523 fixed
poly(age, 3, raw = TRUE)3 -0.00002722 0.00002044 -1.332 -0.00006728 0.00001283 fixed
male 0.05268 0.02603 2.024 0.001673 0.1037 fixed
sibling_count3 -0.05402 0.04017 -1.345 -0.1328 0.0247 fixed
sibling_count4 -0.03311 0.04509 -0.7342 -0.1215 0.05527 fixed
sibling_count5 0.04809 0.05423 0.8869 -0.05819 0.1544 fixed
sibling_count5+ 0.1025 0.05516 1.858 -0.005624 0.2106 fixed
birth_order_nonlinear2 -0.04874 0.03365 -1.449 -0.1147 0.01721 fixed
birth_order_nonlinear3 -0.0439 0.04152 -1.057 -0.1253 0.03749 fixed
birth_order_nonlinear4 -0.07397 0.05271 -1.403 -0.1773 0.02935 fixed
birth_order_nonlinear5 -0.09912 0.06679 -1.484 -0.23 0.03179 fixed
birth_order_nonlinear5+ -0.1016 0.06554 -1.55 -0.2301 0.02684 fixed
sd_(Intercept).mother_pidlink 0.2148 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9667 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.7227 0.4465 -1.619 -1.598 0.1524 fixed
poly(age, 3, raw = TRUE)1 0.004249 0.05062 0.08394 -0.09497 0.1035 fixed
poly(age, 3, raw = TRUE)2 0.001573 0.001807 0.8707 -0.001968 0.005114 fixed
poly(age, 3, raw = TRUE)3 -0.00002548 0.00002047 -1.244 -0.0000656 0.00001465 fixed
male 0.0537 0.02603 2.063 0.00269 0.1047 fixed
count_birth_order2/2 -0.09551 0.05926 -1.612 -0.2117 0.02064 fixed
count_birth_order1/3 -0.0317 0.05239 -0.6052 -0.1344 0.07098 fixed
count_birth_order2/3 -0.1419 0.05797 -2.447 -0.2555 -0.02825 fixed
count_birth_order3/3 -0.1565 0.06373 -2.456 -0.2814 -0.03161 fixed
count_birth_order1/4 -0.1047 0.0652 -1.606 -0.2325 0.02307 fixed
count_birth_order2/4 -0.03028 0.0671 -0.4513 -0.1618 0.1012 fixed
count_birth_order3/4 -0.01813 0.07034 -0.2577 -0.156 0.1197 fixed
count_birth_order4/4 -0.2182 0.07436 -2.935 -0.3639 -0.07247 fixed
count_birth_order1/5 -0.08435 0.08798 -0.9588 -0.2568 0.08809 fixed
count_birth_order2/5 0.0693 0.09733 0.712 -0.1215 0.2601 fixed
count_birth_order3/5 0.05283 0.09288 0.5688 -0.1292 0.2349 fixed
count_birth_order4/5 -0.0298 0.08948 -0.333 -0.2052 0.1456 fixed
count_birth_order5/5 -0.08705 0.09509 -0.9154 -0.2734 0.09932 fixed
count_birth_order1/5+ 0.07337 0.08958 0.8191 -0.1022 0.2489 fixed
count_birth_order2/5+ -0.01039 0.08886 -0.1169 -0.1845 0.1638 fixed
count_birth_order3/5+ -0.04537 0.08811 -0.515 -0.2181 0.1273 fixed
count_birth_order4/5+ 0.1355 0.08621 1.571 -0.0335 0.3044 fixed
count_birth_order5/5+ 0.0005431 0.07922 0.006856 -0.1547 0.1558 fixed
count_birth_order5+/5+ -0.01412 0.05865 -0.2408 -0.1291 0.1008 fixed
sd_(Intercept).mother_pidlink 0.2155 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9659 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 16366 16440 -8172 16344 NA NA NA
12 16364 16444 -8170 16340 3.79 1 0.05156
16 16372 16478 -8170 16340 0.7968 4 0.9389
26 16375 16548 -8161 16323 16.83 10 0.07811

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Agreeableness

birthorder <- birthorder %>% mutate(outcome = big5_agree)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5625 0.1596 -3.524 -0.8753 -0.2497 fixed
poly(age, 3, raw = TRUE)1 0.0273 0.01529 1.786 -0.002658 0.05726 fixed
poly(age, 3, raw = TRUE)2 -0.0002583 0.0004491 -0.5752 -0.001139 0.0006219 fixed
poly(age, 3, raw = TRUE)3 -0.0000008022 0.000004131 -0.1942 -0.000008898 0.000007294 fixed
male 0.06076 0.0169 3.596 0.02764 0.09387 fixed
sibling_count3 -0.02089 0.03345 -0.6245 -0.08646 0.04467 fixed
sibling_count4 0.03056 0.03436 0.8896 -0.03677 0.0979 fixed
sibling_count5 0.007767 0.03558 0.2183 -0.06196 0.0775 fixed
sibling_count5+ 0.005771 0.02805 0.2058 -0.0492 0.06074 fixed
sd_(Intercept).mother_pidlink 0.2511 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9667 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5602 0.1596 -3.509 -0.873 -0.2473 fixed
birth_order 0.004807 0.003507 1.371 -0.002065 0.01168 fixed
poly(age, 3, raw = TRUE)1 0.02588 0.01532 1.689 -0.004146 0.05591 fixed
poly(age, 3, raw = TRUE)2 -0.0002084 0.0004506 -0.4624 -0.001091 0.0006747 fixed
poly(age, 3, raw = TRUE)3 -0.000001256 0.000004144 -0.303 -0.000009377 0.000006866 fixed
male 0.06062 0.0169 3.588 0.02751 0.09373 fixed
sibling_count3 -0.0221 0.03346 -0.6604 -0.08768 0.04349 fixed
sibling_count4 0.02712 0.03445 0.7873 -0.0404 0.09463 fixed
sibling_count5 0.001865 0.03584 0.05204 -0.06837 0.0721 fixed
sibling_count5+ -0.01264 0.03109 -0.4065 -0.07358 0.0483 fixed
sd_(Intercept).mother_pidlink 0.251 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9667 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5594 0.1601 -3.495 -0.8731 -0.2457 fixed
poly(age, 3, raw = TRUE)1 0.02613 0.01533 1.705 -0.003906 0.05617 fixed
poly(age, 3, raw = TRUE)2 -0.0002155 0.0004507 -0.4782 -0.001099 0.0006679 fixed
poly(age, 3, raw = TRUE)3 -0.000001198 0.000004146 -0.2889 -0.000009323 0.000006927 fixed
male 0.0607 0.0169 3.592 0.02758 0.09382 fixed
sibling_count3 -0.01916 0.03398 -0.5638 -0.08576 0.04744 fixed
sibling_count4 0.02892 0.03552 0.814 -0.0407 0.09853 fixed
sibling_count5 0.007183 0.03737 0.1922 -0.06606 0.08043 fixed
sibling_count5+ -0.01117 0.03282 -0.3404 -0.0755 0.05315 fixed
birth_order_nonlinear2 0.008944 0.0247 0.3622 -0.03946 0.05735 fixed
birth_order_nonlinear3 -0.002557 0.02915 -0.08772 -0.0597 0.05458 fixed
birth_order_nonlinear4 0.02067 0.0332 0.6226 -0.0444 0.08574 fixed
birth_order_nonlinear5 -0.001727 0.03775 -0.04573 -0.07572 0.07227 fixed
birth_order_nonlinear5+ 0.04206 0.03104 1.355 -0.01877 0.1029 fixed
sd_(Intercept).mother_pidlink 0.251 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9668 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5441 0.1607 -3.384 -0.8591 -0.229 fixed
poly(age, 3, raw = TRUE)1 0.0261 0.01533 1.703 -0.003943 0.05615 fixed
poly(age, 3, raw = TRUE)2 -0.0002053 0.0004508 -0.4555 -0.001089 0.0006782 fixed
poly(age, 3, raw = TRUE)3 -0.000001373 0.000004146 -0.331 -0.0000095 0.000006754 fixed
male 0.06108 0.0169 3.615 0.02796 0.0942 fixed
count_birth_order2/2 -0.03887 0.048 -0.8099 -0.133 0.0552 fixed
count_birth_order1/3 -0.02833 0.04506 -0.6288 -0.1167 0.05999 fixed
count_birth_order2/3 0.003943 0.05039 0.07825 -0.09482 0.1027 fixed
count_birth_order3/3 -0.1023 0.05645 -1.812 -0.2129 0.008348 fixed
count_birth_order1/4 0.0002992 0.05146 0.005815 -0.1006 0.1012 fixed
count_birth_order2/4 0.0164 0.05412 0.3031 -0.08967 0.1225 fixed
count_birth_order3/4 -0.02745 0.05881 -0.4668 -0.1427 0.08781 fixed
count_birth_order4/4 0.09601 0.06227 1.542 -0.02603 0.218 fixed
count_birth_order1/5 0.03871 0.0584 0.6629 -0.07574 0.1532 fixed
count_birth_order2/5 -0.03562 0.06143 -0.5797 -0.156 0.08479 fixed
count_birth_order3/5 -0.03819 0.06306 -0.6057 -0.1618 0.0854 fixed
count_birth_order4/5 -0.01414 0.0668 -0.2117 -0.1451 0.1168 fixed
count_birth_order5/5 0.01126 0.06829 0.1649 -0.1226 0.1451 fixed
count_birth_order1/5+ -0.08303 0.04717 -1.76 -0.1755 0.009415 fixed
count_birth_order2/5+ -0.001421 0.04866 -0.02921 -0.09679 0.09394 fixed
count_birth_order3/5+ 0.03321 0.04767 0.6967 -0.06022 0.1266 fixed
count_birth_order4/5+ -0.02795 0.04671 -0.5984 -0.1195 0.06361 fixed
count_birth_order5/5+ -0.03953 0.04706 -0.8398 -0.1318 0.05272 fixed
count_birth_order5+/5+ 0.0126 0.03651 0.3452 -0.05896 0.08416 fixed
sd_(Intercept).mother_pidlink 0.2528 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9662 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 39479 39562 -19728 39457 NA NA NA
12 39479 39569 -19727 39455 1.881 1 0.1702
16 39486 39607 -19727 39454 0.9485 4 0.9175
26 39492 39688 -19720 39440 13.71 10 0.1865

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.645 0.4388 -1.47 -1.505 0.215 fixed
poly(age, 3, raw = TRUE)1 0.03354 0.04978 0.6737 -0.06403 0.1311 fixed
poly(age, 3, raw = TRUE)2 -0.0003367 0.001776 -0.1896 -0.003818 0.003144 fixed
poly(age, 3, raw = TRUE)3 -0.0000005582 0.00002011 -0.02776 -0.00003997 0.00003886 fixed
male 0.05795 0.02567 2.258 0.007642 0.1083 fixed
sibling_count3 -0.04282 0.03968 -1.079 -0.1206 0.03495 fixed
sibling_count4 -0.05145 0.04258 -1.208 -0.1349 0.032 fixed
sibling_count5 0.007037 0.04843 0.1453 -0.08788 0.102 fixed
sibling_count5+ -0.05468 0.04243 -1.288 -0.1378 0.02849 fixed
sd_(Intercept).mother_pidlink 0.2632 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9527 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.6472 0.4389 -1.475 -1.507 0.213 fixed
birth_order 0.002615 0.00843 0.3102 -0.01391 0.01914 fixed
poly(age, 3, raw = TRUE)1 0.03343 0.04979 0.6715 -0.06415 0.131 fixed
poly(age, 3, raw = TRUE)2 -0.0003375 0.001776 -0.19 -0.003819 0.003144 fixed
poly(age, 3, raw = TRUE)3 -0.0000004477 0.00002011 -0.02226 -0.00003987 0.00003898 fixed
male 0.05784 0.02567 2.253 0.007523 0.1082 fixed
sibling_count3 -0.04408 0.03989 -1.105 -0.1223 0.0341 fixed
sibling_count4 -0.05442 0.04364 -1.247 -0.14 0.03112 fixed
sibling_count5 0.002162 0.05092 0.04246 -0.09764 0.102 fixed
sibling_count5+ -0.06446 0.05288 -1.219 -0.1681 0.03918 fixed
sd_(Intercept).mother_pidlink 0.2632 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9528 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.6237 0.4395 -1.419 -1.485 0.2378 fixed
poly(age, 3, raw = TRUE)1 0.03241 0.04982 0.6507 -0.06522 0.1301 fixed
poly(age, 3, raw = TRUE)2 -0.0002935 0.001777 -0.1651 -0.003777 0.00319 fixed
poly(age, 3, raw = TRUE)3 -0.000001054 0.00002013 -0.05235 -0.00004051 0.0000384 fixed
male 0.05752 0.02568 2.24 0.007194 0.1078 fixed
sibling_count3 -0.03331 0.04077 -0.8171 -0.1132 0.04659 fixed
sibling_count4 -0.05023 0.04547 -1.105 -0.1394 0.03889 fixed
sibling_count5 -0.003593 0.05371 -0.06689 -0.1089 0.1017 fixed
sibling_count5+ -0.05582 0.05457 -1.023 -0.1628 0.05114 fixed
birth_order_nonlinear2 -0.03669 0.03347 -1.096 -0.1023 0.02892 fixed
birth_order_nonlinear3 -0.04412 0.04123 -1.07 -0.1249 0.03668 fixed
birth_order_nonlinear4 0.02951 0.05091 0.5796 -0.07027 0.1293 fixed
birth_order_nonlinear5 0.03323 0.06362 0.5224 -0.09146 0.1579 fixed
birth_order_nonlinear5+ -0.03071 0.06333 -0.485 -0.1548 0.0934 fixed
sd_(Intercept).mother_pidlink 0.263 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9529 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.6322 0.4405 -1.435 -1.496 0.2312 fixed
poly(age, 3, raw = TRUE)1 0.03344 0.04993 0.6699 -0.06441 0.1313 fixed
poly(age, 3, raw = TRUE)2 -0.0003295 0.001782 -0.1849 -0.003822 0.003163 fixed
poly(age, 3, raw = TRUE)3 -0.0000006666 0.00002018 -0.03303 -0.00004022 0.00003889 fixed
male 0.05848 0.0257 2.276 0.008111 0.1089 fixed
count_birth_order2/2 -0.03959 0.06047 -0.6547 -0.1581 0.07893 fixed
count_birth_order1/3 -0.02909 0.05298 -0.5491 -0.1329 0.07474 fixed
count_birth_order2/3 -0.05869 0.05791 -1.014 -0.1722 0.0548 fixed
count_birth_order3/3 -0.1059 0.06475 -1.636 -0.2328 0.021 fixed
count_birth_order1/4 -0.06212 0.06478 -0.9589 -0.1891 0.06485 fixed
count_birth_order2/4 -0.1076 0.06708 -1.603 -0.239 0.02392 fixed
count_birth_order3/4 -0.0601 0.07093 -0.8474 -0.1991 0.07891 fixed
count_birth_order4/4 -0.0189 0.07374 -0.2563 -0.1634 0.1256 fixed
count_birth_order1/5 0.04298 0.08804 0.4882 -0.1296 0.2155 fixed
count_birth_order2/5 -0.101 0.09431 -1.071 -0.2858 0.08387 fixed
count_birth_order3/5 -0.02089 0.08841 -0.2363 -0.1942 0.1524 fixed
count_birth_order4/5 -0.05274 0.08515 -0.6193 -0.2196 0.1142 fixed
count_birth_order5/5 0.08936 0.08862 1.008 -0.08434 0.2631 fixed
count_birth_order1/5+ -0.1027 0.08735 -1.176 -0.2739 0.06853 fixed
count_birth_order2/5+ -0.0403 0.08651 -0.4659 -0.2099 0.1292 fixed
count_birth_order3/5+ -0.1274 0.08669 -1.47 -0.2973 0.04252 fixed
count_birth_order4/5+ 0.03782 0.08169 0.463 -0.1223 0.1979 fixed
count_birth_order5/5+ -0.0659 0.0772 -0.8536 -0.2172 0.08541 fixed
count_birth_order5+/5+ -0.08752 0.05776 -1.515 -0.2007 0.02568 fixed
sd_(Intercept).mother_pidlink 0.2633 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9531 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 16672 16745 -8325 16650 NA NA NA
12 16673 16754 -8325 16649 0.09659 1 0.756
16 16677 16784 -8323 16645 3.955 4 0.4121
26 16692 16866 -8320 16640 5.719 10 0.8383

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.6188 0.4369 -1.416 -1.475 0.2376 fixed
poly(age, 3, raw = TRUE)1 0.03119 0.04959 0.629 -0.06601 0.1284 fixed
poly(age, 3, raw = TRUE)2 -0.0002451 0.00177 -0.1385 -0.003714 0.003224 fixed
poly(age, 3, raw = TRUE)3 -0.000001585 0.00002004 -0.07908 -0.00004086 0.00003769 fixed
male 0.05768 0.02554 2.259 0.007634 0.1077 fixed
sibling_count3 -0.05156 0.04288 -1.202 -0.1356 0.03248 fixed
sibling_count4 -0.02386 0.04507 -0.5294 -0.1122 0.06447 fixed
sibling_count5 -0.03461 0.04805 -0.7203 -0.1288 0.05956 fixed
sibling_count5+ -0.0736 0.04215 -1.746 -0.1562 0.009005 fixed
sd_(Intercept).mother_pidlink 0.2616 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.952 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.6268 0.437 -1.434 -1.483 0.2296 fixed
birth_order 0.007124 0.007328 0.9722 -0.007238 0.02149 fixed
poly(age, 3, raw = TRUE)1 0.03115 0.04959 0.6282 -0.06605 0.1283 fixed
poly(age, 3, raw = TRUE)2 -0.0002563 0.00177 -0.1448 -0.003725 0.003212 fixed
poly(age, 3, raw = TRUE)3 -0.000001183 0.00002005 -0.05902 -0.00004047 0.0000381 fixed
male 0.05745 0.02554 2.249 0.007393 0.1075 fixed
sibling_count3 -0.05498 0.04302 -1.278 -0.1393 0.02934 fixed
sibling_count4 -0.03161 0.04577 -0.6907 -0.1213 0.05809 fixed
sibling_count5 -0.04685 0.04967 -0.9433 -0.1442 0.0505 fixed
sibling_count5+ -0.0996 0.04991 -1.995 -0.1974 -0.001771 fixed
sd_(Intercept).mother_pidlink 0.2614 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9521 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.6006 0.4376 -1.373 -1.458 0.2571 fixed
poly(age, 3, raw = TRUE)1 0.03018 0.04962 0.6083 -0.06707 0.1274 fixed
poly(age, 3, raw = TRUE)2 -0.0002201 0.001771 -0.1243 -0.003691 0.003251 fixed
poly(age, 3, raw = TRUE)3 -0.000001571 0.00002006 -0.07831 -0.00004089 0.00003775 fixed
male 0.05734 0.02555 2.244 0.007266 0.1074 fixed
sibling_count3 -0.04586 0.04386 -1.046 -0.1318 0.0401 fixed
sibling_count4 -0.0244 0.04748 -0.5138 -0.1175 0.06867 fixed
sibling_count5 -0.04551 0.05221 -0.8717 -0.1478 0.05682 fixed
sibling_count5+ -0.1009 0.05158 -1.957 -0.202 0.0001439 fixed
birth_order_nonlinear2 -0.02657 0.03404 -0.7807 -0.09328 0.04014 fixed
birth_order_nonlinear3 -0.02837 0.04103 -0.6913 -0.1088 0.05206 fixed
birth_order_nonlinear4 0.01991 0.04932 0.4036 -0.07676 0.1166 fixed
birth_order_nonlinear5 0.04047 0.06051 0.6688 -0.07813 0.1591 fixed
birth_order_nonlinear5+ 0.04104 0.05641 0.7276 -0.06952 0.1516 fixed
sd_(Intercept).mother_pidlink 0.2608 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9524 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.6018 0.4385 -1.372 -1.461 0.2578 fixed
poly(age, 3, raw = TRUE)1 0.03204 0.04971 0.6445 -0.06539 0.1295 fixed
poly(age, 3, raw = TRUE)2 -0.0002896 0.001775 -0.1632 -0.003768 0.003189 fixed
poly(age, 3, raw = TRUE)3 -0.0000007273 0.0000201 -0.03618 -0.00004013 0.00003868 fixed
male 0.05762 0.02557 2.253 0.007493 0.1077 fixed
count_birth_order2/2 -0.07001 0.0661 -1.059 -0.1996 0.05955 fixed
count_birth_order1/3 -0.05588 0.05731 -0.9752 -0.1682 0.05643 fixed
count_birth_order2/3 -0.06772 0.06217 -1.089 -0.1896 0.05414 fixed
count_birth_order3/3 -0.127 0.07 -1.814 -0.2642 0.0102 fixed
count_birth_order1/4 -0.05227 0.06778 -0.7711 -0.1851 0.08058 fixed
count_birth_order2/4 -0.04976 0.06909 -0.7201 -0.1852 0.08567 fixed
count_birth_order3/4 -0.07543 0.07588 -0.9941 -0.2241 0.07329 fixed
count_birth_order4/4 -0.01413 0.07823 -0.1806 -0.1675 0.1392 fixed
count_birth_order1/5 -0.03582 0.08047 -0.4451 -0.1935 0.1219 fixed
count_birth_order2/5 -0.1593 0.08648 -1.842 -0.3288 0.01021 fixed
count_birth_order3/5 -0.03115 0.08411 -0.3704 -0.196 0.1337 fixed
count_birth_order4/5 -0.06292 0.0872 -0.7215 -0.2338 0.108 fixed
count_birth_order5/5 -0.0179 0.08745 -0.2047 -0.1893 0.1535 fixed
count_birth_order1/5+ -0.1703 0.07649 -2.227 -0.3203 -0.02043 fixed
count_birth_order2/5+ -0.1037 0.07995 -1.297 -0.2604 0.05303 fixed
count_birth_order3/5+ -0.1335 0.07845 -1.701 -0.2873 0.02028 fixed
count_birth_order4/5+ -0.08556 0.07552 -1.133 -0.2336 0.06246 fixed
count_birth_order5/5+ -0.07678 0.07763 -0.989 -0.2289 0.07538 fixed
count_birth_order5+/5+ -0.07471 0.05656 -1.321 -0.1856 0.03614 fixed
sd_(Intercept).mother_pidlink 0.2602 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9531 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 16802 16876 -8390 16780 NA NA NA
12 16803 16884 -8390 16779 0.9475 1 0.3304
16 16810 16917 -8389 16778 1.889 4 0.7562
26 16825 17000 -8387 16773 4.132 10 0.9412

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5969 0.4431 -1.347 -1.465 0.2716 fixed
poly(age, 3, raw = TRUE)1 0.02716 0.0503 0.5399 -0.07142 0.1257 fixed
poly(age, 3, raw = TRUE)2 -0.00008155 0.001795 -0.04543 -0.0036 0.003437 fixed
poly(age, 3, raw = TRUE)3 -0.000003671 0.00002033 -0.1805 -0.00004353 0.00003618 fixed
male 0.0572 0.02589 2.21 0.00646 0.1079 fixed
sibling_count3 -0.04753 0.03907 -1.217 -0.1241 0.02904 fixed
sibling_count4 -0.06009 0.04218 -1.425 -0.1428 0.02258 fixed
sibling_count5 0.04759 0.04945 0.9624 -0.04933 0.1445 fixed
sibling_count5+ -0.0633 0.04274 -1.481 -0.1471 0.02047 fixed
sd_(Intercept).mother_pidlink 0.2586 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9519 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5982 0.4432 -1.35 -1.467 0.2704 fixed
birth_order 0.001925 0.008673 0.222 -0.01507 0.01892 fixed
poly(age, 3, raw = TRUE)1 0.02705 0.0503 0.5377 -0.07154 0.1256 fixed
poly(age, 3, raw = TRUE)2 -0.00008128 0.001795 -0.04528 -0.0036 0.003437 fixed
poly(age, 3, raw = TRUE)3 -0.000003598 0.00002034 -0.1769 -0.00004346 0.00003627 fixed
male 0.05715 0.02589 2.207 0.006407 0.1079 fixed
sibling_count3 -0.04846 0.03929 -1.233 -0.1255 0.02856 fixed
sibling_count4 -0.06225 0.04329 -1.438 -0.1471 0.0226 fixed
sibling_count5 0.04417 0.05181 0.8525 -0.05738 0.1457 fixed
sibling_count5+ -0.07045 0.05352 -1.316 -0.1753 0.03444 fixed
sd_(Intercept).mother_pidlink 0.2587 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9519 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5775 0.4438 -1.301 -1.447 0.2924 fixed
poly(age, 3, raw = TRUE)1 0.02592 0.05033 0.515 -0.07272 0.1246 fixed
poly(age, 3, raw = TRUE)2 -0.00003914 0.001796 -0.02179 -0.00356 0.003481 fixed
poly(age, 3, raw = TRUE)3 -0.000004064 0.00002035 -0.1997 -0.00004396 0.00003583 fixed
male 0.0566 0.0259 2.186 0.005845 0.1074 fixed
sibling_count3 -0.03684 0.0402 -0.9166 -0.1156 0.04194 fixed
sibling_count4 -0.05666 0.04517 -1.254 -0.1452 0.03188 fixed
sibling_count5 0.04475 0.0544 0.8225 -0.06188 0.1514 fixed
sibling_count5+ -0.06844 0.05532 -1.237 -0.1769 0.03998 fixed
birth_order_nonlinear2 -0.02562 0.03336 -0.768 -0.09101 0.03976 fixed
birth_order_nonlinear3 -0.04947 0.0412 -1.201 -0.1302 0.03128 fixed
birth_order_nonlinear4 0.0286 0.05233 0.5465 -0.07397 0.1312 fixed
birth_order_nonlinear5 0.007749 0.06629 0.1169 -0.1222 0.1377 fixed
birth_order_nonlinear5+ -0.00116 0.06527 -0.01777 -0.1291 0.1268 fixed
sd_(Intercept).mother_pidlink 0.2588 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.952 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5521 0.4449 -1.241 -1.424 0.3198 fixed
poly(age, 3, raw = TRUE)1 0.02312 0.05045 0.4584 -0.07575 0.122 fixed
poly(age, 3, raw = TRUE)2 0.00006625 0.001801 0.03679 -0.003463 0.003596 fixed
poly(age, 3, raw = TRUE)3 -0.000005314 0.00002041 -0.2604 -0.00004531 0.00003468 fixed
male 0.05695 0.02592 2.197 0.006146 0.1078 fixed
count_birth_order2/2 -0.03262 0.05883 -0.5544 -0.1479 0.08269 fixed
count_birth_order1/3 -0.03163 0.05223 -0.6057 -0.134 0.07073 fixed
count_birth_order2/3 -0.05113 0.05778 -0.885 -0.1644 0.06211 fixed
count_birth_order3/3 -0.1218 0.0635 -1.918 -0.2462 0.002663 fixed
count_birth_order1/4 -0.07897 0.06499 -1.215 -0.2063 0.0484 fixed
count_birth_order2/4 -0.1058 0.06687 -1.583 -0.2369 0.02521 fixed
count_birth_order3/4 -0.07093 0.07008 -1.012 -0.2083 0.06642 fixed
count_birth_order4/4 -0.01687 0.07408 -0.2277 -0.1621 0.1283 fixed
count_birth_order1/5 0.05054 0.08768 0.5764 -0.1213 0.2224 fixed
count_birth_order2/5 -0.01504 0.09697 -0.1551 -0.2051 0.175 fixed
count_birth_order3/5 -0.0003237 0.09252 -0.003499 -0.1817 0.181 fixed
count_birth_order4/5 -0.02 0.08914 -0.2244 -0.1947 0.1547 fixed
count_birth_order5/5 0.1681 0.09472 1.774 -0.0176 0.3537 fixed
count_birth_order1/5+ -0.0804 0.08927 -0.9006 -0.2554 0.09458 fixed
count_birth_order2/5+ -0.05915 0.08855 -0.668 -0.2327 0.1144 fixed
count_birth_order3/5+ -0.1176 0.08777 -1.34 -0.2897 0.0544 fixed
count_birth_order4/5+ 0.02036 0.08587 0.2372 -0.1479 0.1887 fixed
count_birth_order5/5+ -0.1374 0.07891 -1.741 -0.2921 0.01728 fixed
count_birth_order5+/5+ -0.07222 0.05867 -1.231 -0.1872 0.04277 fixed
sd_(Intercept).mother_pidlink 0.2592 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9522 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 16314 16387 -8146 16292 NA NA NA
12 16316 16396 -8146 16292 0.0492 1 0.8245
16 16321 16428 -8144 16289 2.826 4 0.5874
26 16334 16508 -8141 16282 6.65 10 0.758

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Openness

birthorder <- birthorder %>% mutate(outcome = big5_open)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.4167 0.1589 2.623 0.1053 0.728 fixed
poly(age, 3, raw = TRUE)1 -0.03942 0.01522 -2.59 -0.06925 -0.009592 fixed
poly(age, 3, raw = TRUE)2 0.001142 0.0004473 2.552 0.0002649 0.002018 fixed
poly(age, 3, raw = TRUE)3 -0.00001113 0.000004115 -2.705 -0.0000192 -0.000003067 fixed
male 0.1705 0.01678 10.16 0.1377 0.2034 fixed
sibling_count3 -0.0164 0.03344 -0.4905 -0.08194 0.04913 fixed
sibling_count4 -0.04541 0.03438 -1.321 -0.1128 0.02197 fixed
sibling_count5 -0.004679 0.03564 -0.1313 -0.07454 0.06518 fixed
sibling_count5+ -0.07108 0.02806 -2.533 -0.1261 -0.01608 fixed
sd_(Intercept).mother_pidlink 0.2818 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9529 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.4136 0.1588 2.604 0.1023 0.7249 fixed
birth_order -0.006737 0.003501 -1.924 -0.0136 0.0001249 fixed
poly(age, 3, raw = TRUE)1 -0.03742 0.01525 -2.453 -0.06731 -0.007525 fixed
poly(age, 3, raw = TRUE)2 0.001071 0.0004488 2.386 0.000191 0.00195 fixed
poly(age, 3, raw = TRUE)3 -0.00001049 0.000004129 -2.54 -0.00001858 -0.000002394 fixed
male 0.1707 0.01678 10.18 0.1378 0.2036 fixed
sibling_count3 -0.01476 0.03344 -0.4413 -0.0803 0.05079 fixed
sibling_count4 -0.04063 0.03446 -1.179 -0.1082 0.02691 fixed
sibling_count5 0.003553 0.03589 0.099 -0.06679 0.0739 fixed
sibling_count5+ -0.04537 0.03107 -1.46 -0.1063 0.01553 fixed
sd_(Intercept).mother_pidlink 0.2814 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9529 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.4067 0.1593 2.554 0.09454 0.7189 fixed
poly(age, 3, raw = TRUE)1 -0.03765 0.01526 -2.467 -0.06755 -0.007742 fixed
poly(age, 3, raw = TRUE)2 0.001092 0.0004489 2.432 0.0002121 0.001972 fixed
poly(age, 3, raw = TRUE)3 -0.00001079 0.00000413 -2.613 -0.00001889 -0.000002696 fixed
male 0.1708 0.01678 10.18 0.1379 0.2036 fixed
sibling_count3 -0.004799 0.03395 -0.1414 -0.07133 0.06173 fixed
sibling_count4 -0.02479 0.0355 -0.6983 -0.09438 0.04479 fixed
sibling_count5 0.01456 0.03739 0.3894 -0.05872 0.08783 fixed
sibling_count5+ -0.04095 0.03276 -1.25 -0.1051 0.02325 fixed
birth_order_nonlinear2 -0.01379 0.02447 -0.5636 -0.06176 0.03417 fixed
birth_order_nonlinear3 -0.0592 0.02888 -2.049 -0.1158 -0.002584 fixed
birth_order_nonlinear4 -0.05691 0.0329 -1.73 -0.1214 0.007566 fixed
birth_order_nonlinear5 -0.007878 0.03741 -0.2106 -0.08121 0.06545 fixed
birth_order_nonlinear5+ -0.04945 0.03085 -1.603 -0.1099 0.01101 fixed
sd_(Intercept).mother_pidlink 0.2817 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9528 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.4334 0.16 2.709 0.1198 0.7469 fixed
poly(age, 3, raw = TRUE)1 -0.03779 0.01526 -2.477 -0.0677 -0.007886 fixed
poly(age, 3, raw = TRUE)2 0.001099 0.0004489 2.448 0.000219 0.001979 fixed
poly(age, 3, raw = TRUE)3 -0.00001087 0.000004131 -2.633 -0.00001897 -0.000002778 fixed
male 0.1706 0.01678 10.16 0.1377 0.2034 fixed
count_birth_order2/2 -0.08334 0.04757 -1.752 -0.1766 0.009892 fixed
count_birth_order1/3 -0.0571 0.04482 -1.274 -0.1449 0.03074 fixed
count_birth_order2/3 -0.02348 0.05011 -0.4685 -0.1217 0.07474 fixed
count_birth_order3/3 -0.0652 0.05612 -1.162 -0.1752 0.0448 fixed
count_birth_order1/4 -0.07077 0.05117 -1.383 -0.1711 0.02953 fixed
count_birth_order2/4 -0.009176 0.05382 -0.1705 -0.1147 0.0963 fixed
count_birth_order3/4 -0.1791 0.05847 -3.062 -0.2936 -0.06445 fixed
count_birth_order4/4 -0.07479 0.0619 -1.208 -0.1961 0.04654 fixed
count_birth_order1/5 0.0005729 0.05806 0.009867 -0.1132 0.1144 fixed
count_birth_order2/5 -0.06346 0.06107 -1.039 -0.1832 0.05625 fixed
count_birth_order3/5 -0.07172 0.06269 -1.144 -0.1946 0.05114 fixed
count_birth_order4/5 -0.02146 0.0664 -0.3232 -0.1516 0.1087 fixed
count_birth_order5/5 -0.03514 0.06788 -0.5177 -0.1682 0.09789 fixed
count_birth_order1/5+ -0.06634 0.04689 -1.415 -0.1582 0.02555 fixed
count_birth_order2/5+ -0.07713 0.04836 -1.595 -0.1719 0.01766 fixed
count_birth_order3/5+ -0.1038 0.04738 -2.19 -0.1966 -0.01091 fixed
count_birth_order4/5+ -0.1541 0.04643 -3.319 -0.2451 -0.06311 fixed
count_birth_order5/5+ -0.06927 0.04678 -1.481 -0.161 0.02241 fixed
count_birth_order5+/5+ -0.1164 0.0364 -3.198 -0.1877 -0.04506 fixed
sd_(Intercept).mother_pidlink 0.2812 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.953 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 39311 39394 -19645 39289 NA NA NA
12 39310 39400 -19643 39286 3.706 1 0.05421
16 39315 39435 -19641 39283 3.124 4 0.5373
26 39324 39521 -19636 39272 10.07 10 0.4341

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.833 0.4144 -2.01 -1.645 -0.02067 fixed
poly(age, 3, raw = TRUE)1 0.1105 0.04703 2.349 0.0183 0.2026 fixed
poly(age, 3, raw = TRUE)2 -0.004051 0.001678 -2.414 -0.00734 -0.0007618 fixed
poly(age, 3, raw = TRUE)3 0.00004751 0.000019 2.5 0.00001026 0.00008475 fixed
male 0.1386 0.02423 5.719 0.09108 0.1861 fixed
sibling_count3 -0.09714 0.03779 -2.571 -0.1712 -0.02308 fixed
sibling_count4 -0.08943 0.04063 -2.201 -0.1691 -0.009798 fixed
sibling_count5 -0.1136 0.04629 -2.453 -0.2043 -0.02283 fixed
sibling_count5+ -0.1441 0.04061 -3.548 -0.2237 -0.06447 fixed
sd_(Intercept).mother_pidlink 0.2961 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8871 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.8361 0.4145 -2.017 -1.649 -0.02365 fixed
birth_order 0.003801 0.007987 0.4758 -0.01185 0.01946 fixed
poly(age, 3, raw = TRUE)1 0.1103 0.04703 2.345 0.01813 0.2025 fixed
poly(age, 3, raw = TRUE)2 -0.004051 0.001678 -2.414 -0.007341 -0.0007619 fixed
poly(age, 3, raw = TRUE)3 0.00004766 0.00001901 2.507 0.00001041 0.00008491 fixed
male 0.1384 0.02423 5.711 0.0909 0.1859 fixed
sibling_count3 -0.09898 0.03799 -2.606 -0.1734 -0.02453 fixed
sibling_count4 -0.09377 0.04164 -2.252 -0.1754 -0.01216 fixed
sibling_count5 -0.1207 0.04866 -2.481 -0.2161 -0.02533 fixed
sibling_count5+ -0.1583 0.0505 -3.136 -0.2573 -0.05937 fixed
sd_(Intercept).mother_pidlink 0.296 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8872 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.8633 0.4151 -2.08 -1.677 -0.04975 fixed
poly(age, 3, raw = TRUE)1 0.113 0.04705 2.402 0.02079 0.2052 fixed
poly(age, 3, raw = TRUE)2 -0.00414 0.001679 -2.466 -0.007431 -0.0008496 fixed
poly(age, 3, raw = TRUE)3 0.0000485 0.00001902 2.55 0.00001123 0.00008577 fixed
male 0.1388 0.02423 5.727 0.09127 0.1862 fixed
sibling_count3 -0.0928 0.03879 -2.392 -0.1688 -0.01678 fixed
sibling_count4 -0.09456 0.04332 -2.183 -0.1795 -0.009643 fixed
sibling_count5 -0.1389 0.05122 -2.712 -0.2393 -0.03852 fixed
sibling_count5+ -0.146 0.05205 -2.806 -0.248 -0.04402 fixed
birth_order_nonlinear2 0.02394 0.03144 0.7614 -0.03769 0.08556 fixed
birth_order_nonlinear3 -0.01775 0.03877 -0.4578 -0.09373 0.05823 fixed
birth_order_nonlinear4 0.05292 0.0479 1.105 -0.04096 0.1468 fixed
birth_order_nonlinear5 0.106 0.05983 1.771 -0.01128 0.2233 fixed
birth_order_nonlinear5+ -0.03825 0.05984 -0.6391 -0.1555 0.07904 fixed
sd_(Intercept).mother_pidlink 0.2962 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8869 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.8675 0.416 -2.085 -1.683 -0.05216 fixed
poly(age, 3, raw = TRUE)1 0.1149 0.04715 2.436 0.02247 0.2073 fixed
poly(age, 3, raw = TRUE)2 -0.004197 0.001683 -2.494 -0.007495 -0.0008982 fixed
poly(age, 3, raw = TRUE)3 0.00004901 0.00001906 2.571 0.00001164 0.00008637 fixed
male 0.1395 0.02425 5.752 0.09196 0.187 fixed
count_birth_order2/2 -0.01995 0.05685 -0.351 -0.1314 0.09147 fixed
count_birth_order1/3 -0.137 0.05008 -2.736 -0.2352 -0.03889 fixed
count_birth_order2/3 -0.07656 0.05472 -1.399 -0.1838 0.03069 fixed
count_birth_order3/3 -0.0773 0.06117 -1.264 -0.1972 0.04259 fixed
count_birth_order1/4 -0.07394 0.06123 -1.208 -0.194 0.04607 fixed
count_birth_order2/4 -0.1317 0.06338 -2.078 -0.2559 -0.007483 fixed
count_birth_order3/4 -0.1464 0.06698 -2.186 -0.2777 -0.01516 fixed
count_birth_order4/4 -0.02691 0.06964 -0.3864 -0.1634 0.1096 fixed
count_birth_order1/5 -0.1685 0.08317 -2.026 -0.3315 -0.005496 fixed
count_birth_order2/5 -0.05147 0.08905 -0.578 -0.226 0.1231 fixed
count_birth_order3/5 -0.2027 0.08347 -2.428 -0.3663 -0.03908 fixed
count_birth_order4/5 -0.1164 0.08041 -1.448 -0.274 0.04115 fixed
count_birth_order5/5 -0.05156 0.08367 -0.6162 -0.2156 0.1124 fixed
count_birth_order1/5+ -0.1743 0.08252 -2.112 -0.336 -0.01253 fixed
count_birth_order2/5+ -0.06251 0.0817 -0.7651 -0.2226 0.09762 fixed
count_birth_order3/5+ -0.2153 0.08185 -2.631 -0.3757 -0.05489 fixed
count_birth_order4/5+ -0.1316 0.07712 -1.707 -0.2828 0.01951 fixed
count_birth_order5/5+ -0.0526 0.07288 -0.7217 -0.1954 0.09024 fixed
count_birth_order5+/5+ -0.1993 0.05481 -3.636 -0.3067 -0.09185 fixed
sd_(Intercept).mother_pidlink 0.2959 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8872 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 16002 16076 -7990 15980 NA NA NA
12 16004 16084 -7990 15980 0.2274 1 0.6335
16 16005 16112 -7986 15973 7.314 4 0.1202
26 16018 16192 -7983 15966 6.635 10 0.7594

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.7636 0.4132 -1.848 -1.573 0.04634 fixed
poly(age, 3, raw = TRUE)1 0.1043 0.04691 2.224 0.01237 0.1963 fixed
poly(age, 3, raw = TRUE)2 -0.003837 0.001674 -2.292 -0.007119 -0.0005553 fixed
poly(age, 3, raw = TRUE)3 0.00004507 0.00001896 2.376 0.000007898 0.00008223 fixed
male 0.1403 0.02414 5.814 0.09303 0.1876 fixed
sibling_count3 -0.1248 0.04088 -3.052 -0.2049 -0.04466 fixed
sibling_count4 -0.1013 0.04303 -2.354 -0.1856 -0.01694 fixed
sibling_count5 -0.1191 0.04596 -2.592 -0.2092 -0.02906 fixed
sibling_count5+ -0.134 0.0403 -3.325 -0.213 -0.055 fixed
sd_(Intercept).mother_pidlink 0.2966 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8872 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.7645 0.4133 -1.85 -1.575 0.04561 fixed
birth_order 0.0008314 0.006965 0.1194 -0.01282 0.01448 fixed
poly(age, 3, raw = TRUE)1 0.1043 0.04691 2.223 0.01236 0.1963 fixed
poly(age, 3, raw = TRUE)2 -0.003838 0.001674 -2.292 -0.00712 -0.0005563 fixed
poly(age, 3, raw = TRUE)3 0.00004511 0.00001897 2.378 0.000007934 0.00008229 fixed
male 0.1403 0.02414 5.812 0.09299 0.1876 fixed
sibling_count3 -0.1252 0.04102 -3.052 -0.2056 -0.04478 fixed
sibling_count4 -0.1022 0.04371 -2.338 -0.1879 -0.01652 fixed
sibling_count5 -0.1206 0.04751 -2.538 -0.2137 -0.02745 fixed
sibling_count5+ -0.137 0.04768 -2.874 -0.2305 -0.04357 fixed
sd_(Intercept).mother_pidlink 0.2967 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8873 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.7658 0.4139 -1.85 -1.577 0.04536 fixed
poly(age, 3, raw = TRUE)1 0.1043 0.04693 2.221 0.01227 0.1963 fixed
poly(age, 3, raw = TRUE)2 -0.00383 0.001675 -2.286 -0.007114 -0.0005466 fixed
poly(age, 3, raw = TRUE)3 0.0000449 0.00001898 2.366 0.000007699 0.0000821 fixed
male 0.1403 0.02414 5.81 0.09296 0.1876 fixed
sibling_count3 -0.1194 0.04179 -2.857 -0.2013 -0.0375 fixed
sibling_count4 -0.1057 0.04529 -2.333 -0.1944 -0.01689 fixed
sibling_count5 -0.1304 0.04986 -2.614 -0.2281 -0.03263 fixed
sibling_count5+ -0.126 0.04921 -2.561 -0.2225 -0.02959 fixed
birth_order_nonlinear2 0.006196 0.03202 0.1935 -0.05657 0.06896 fixed
birth_order_nonlinear3 -0.02327 0.03865 -0.6022 -0.09902 0.05247 fixed
birth_order_nonlinear4 0.05037 0.04648 1.084 -0.04073 0.1415 fixed
birth_order_nonlinear5 0.04243 0.05703 0.744 -0.06934 0.1542 fixed
birth_order_nonlinear5+ -0.04304 0.05343 -0.8055 -0.1478 0.06169 fixed
sd_(Intercept).mother_pidlink 0.2968 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8872 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.7433 0.4145 -1.793 -1.556 0.06915 fixed
poly(age, 3, raw = TRUE)1 0.1031 0.04699 2.194 0.011 0.1952 fixed
poly(age, 3, raw = TRUE)2 -0.003777 0.001678 -2.251 -0.007065 -0.0004884 fixed
poly(age, 3, raw = TRUE)3 0.00004418 0.00001901 2.324 0.000006919 0.00008145 fixed
male 0.1401 0.02415 5.801 0.09278 0.1875 fixed
count_birth_order2/2 -0.03849 0.0622 -0.6188 -0.1604 0.08342 fixed
count_birth_order1/3 -0.1582 0.05423 -2.917 -0.2645 -0.05192 fixed
count_birth_order2/3 -0.1277 0.05882 -2.171 -0.243 -0.01239 fixed
count_birth_order3/3 -0.1121 0.0662 -1.693 -0.2418 0.01764 fixed
count_birth_order1/4 -0.1371 0.06414 -2.138 -0.2628 -0.01142 fixed
count_birth_order2/4 -0.1088 0.06536 -1.665 -0.2369 0.01928 fixed
count_birth_order3/4 -0.2025 0.07174 -2.823 -0.3431 -0.06191 fixed
count_birth_order4/4 0.008894 0.07396 0.1203 -0.1361 0.1539 fixed
count_birth_order1/5 -0.1709 0.07612 -2.245 -0.3201 -0.02171 fixed
count_birth_order2/5 -0.161 0.08176 -1.969 -0.3212 -0.0007141 fixed
count_birth_order3/5 -0.1594 0.07951 -2.004 -0.3152 -0.003521 fixed
count_birth_order4/5 -0.03537 0.08242 -0.4291 -0.1969 0.1262 fixed
count_birth_order5/5 -0.1204 0.08264 -1.457 -0.2823 0.0416 fixed
count_birth_order1/5+ -0.084 0.07236 -1.161 -0.2258 0.05781 fixed
count_birth_order2/5+ -0.08335 0.0756 -1.103 -0.2315 0.06482 fixed
count_birth_order3/5+ -0.17 0.07416 -2.293 -0.3154 -0.0247 fixed
count_birth_order4/5+ -0.201 0.07138 -2.816 -0.3409 -0.06113 fixed
count_birth_order5/5+ -0.08702 0.07335 -1.186 -0.2308 0.05674 fixed
count_birth_order5+/5+ -0.1847 0.05372 -3.439 -0.29 -0.07944 fixed
sd_(Intercept).mother_pidlink 0.2977 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8869 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 16143 16217 -8061 16121 NA NA NA
12 16145 16225 -8061 16121 0.01427 1 0.9049
16 16149 16256 -8058 16117 4.361 4 0.3594
26 16158 16332 -8053 16106 10.45 10 0.4015

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.7587 0.4191 -1.81 -1.58 0.06275 fixed
poly(age, 3, raw = TRUE)1 0.1017 0.04758 2.138 0.00846 0.195 fixed
poly(age, 3, raw = TRUE)2 -0.003723 0.001698 -2.192 -0.007051 -0.0003944 fixed
poly(age, 3, raw = TRUE)3 0.00004357 0.00001924 2.264 0.000005855 0.00008128 fixed
male 0.1344 0.02447 5.493 0.08644 0.1823 fixed
sibling_count3 -0.1099 0.03728 -2.949 -0.183 -0.03687 fixed
sibling_count4 -0.09092 0.04034 -2.254 -0.17 -0.01184 fixed
sibling_count5 -0.09754 0.04741 -2.057 -0.1905 -0.004609 fixed
sibling_count5+ -0.1308 0.04103 -3.188 -0.2112 -0.05037 fixed
sd_(Intercept).mother_pidlink 0.2968 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8863 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.7619 0.4192 -1.818 -1.583 0.05964 fixed
birth_order 0.004791 0.008228 0.5822 -0.01134 0.02092 fixed
poly(age, 3, raw = TRUE)1 0.1014 0.04758 2.131 0.008162 0.1947 fixed
poly(age, 3, raw = TRUE)2 -0.003721 0.001698 -2.191 -0.00705 -0.0003926 fixed
poly(age, 3, raw = TRUE)3 0.00004374 0.00001924 2.273 0.000006022 0.00008146 fixed
male 0.1343 0.02447 5.487 0.08631 0.1822 fixed
sibling_count3 -0.1123 0.0375 -2.994 -0.1858 -0.03878 fixed
sibling_count4 -0.09632 0.0414 -2.327 -0.1775 -0.01518 fixed
sibling_count5 -0.1061 0.04965 -2.137 -0.2034 -0.008801 fixed
sibling_count5+ -0.1486 0.05122 -2.902 -0.249 -0.04824 fixed
sd_(Intercept).mother_pidlink 0.2968 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8863 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.7727 0.4197 -1.841 -1.595 0.04996 fixed
poly(age, 3, raw = TRUE)1 0.1028 0.0476 2.16 0.009528 0.1961 fixed
poly(age, 3, raw = TRUE)2 -0.003761 0.001699 -2.214 -0.007092 -0.0004311 fixed
poly(age, 3, raw = TRUE)3 0.00004398 0.00001926 2.284 0.000006237 0.00008172 fixed
male 0.1339 0.02447 5.473 0.08598 0.1819 fixed
sibling_count3 -0.1077 0.03832 -2.811 -0.1828 -0.03261 fixed
sibling_count4 -0.1011 0.04312 -2.345 -0.1856 -0.0166 fixed
sibling_count5 -0.1219 0.05202 -2.344 -0.2239 -0.01998 fixed
sibling_count5+ -0.1329 0.05286 -2.514 -0.2365 -0.02927 fixed
birth_order_nonlinear2 0.01259 0.03137 0.4012 -0.0489 0.07407 fixed
birth_order_nonlinear3 -0.007885 0.03878 -0.2033 -0.08389 0.06812 fixed
birth_order_nonlinear4 0.06473 0.0493 1.313 -0.03189 0.1613 fixed
birth_order_nonlinear5 0.07644 0.06242 1.225 -0.04591 0.1988 fixed
birth_order_nonlinear5+ -0.03802 0.06175 -0.6156 -0.159 0.08301 fixed
sd_(Intercept).mother_pidlink 0.2956 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.8866 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.7574 0.4207 -1.8 -1.582 0.06723 fixed
poly(age, 3, raw = TRUE)1 0.1027 0.04771 2.152 0.009185 0.1962 fixed
poly(age, 3, raw = TRUE)2 -0.003749 0.001703 -2.201 -0.007088 -0.0004106 fixed
poly(age, 3, raw = TRUE)3 0.00004374 0.00001931 2.266 0.0000059 0.00008159 fixed
male 0.1337 0.02449 5.46 0.08574 0.1818 fixed
count_birth_order2/2 -0.03314 0.05536 -0.5986 -0.1417 0.07537 fixed
count_birth_order1/3 -0.1392 0.04945 -2.815 -0.2361 -0.04228 fixed
count_birth_order2/3 -0.1234 0.05468 -2.257 -0.2306 -0.01625 fixed
count_birth_order3/3 -0.08378 0.06007 -1.395 -0.2015 0.03395 fixed
count_birth_order1/4 -0.1067 0.06151 -1.735 -0.2273 0.01385 fixed
count_birth_order2/4 -0.1168 0.06327 -1.846 -0.2408 0.007236 fixed
count_birth_order3/4 -0.1429 0.06628 -2.156 -0.2728 -0.01301 fixed
count_birth_order4/4 -0.02894 0.07005 -0.4132 -0.1662 0.1084 fixed
count_birth_order1/5 -0.2023 0.08297 -2.438 -0.3649 -0.03966 fixed
count_birth_order2/5 -0.03606 0.0917 -0.3932 -0.2158 0.1437 fixed
count_birth_order3/5 -0.1885 0.08748 -2.154 -0.3599 -0.01701 fixed
count_birth_order4/5 -0.05182 0.08429 -0.6147 -0.217 0.1134 fixed
count_birth_order5/5 -0.04479 0.08956 -0.5001 -0.2203 0.1307 fixed
count_birth_order1/5+ -0.1186 0.08445 -1.404 -0.2841 0.04694 fixed
count_birth_order2/5+ -0.06131 0.08375 -0.7321 -0.2255 0.1028 fixed
count_birth_order3/5+ -0.1921 0.08298 -2.315 -0.3548 -0.02946 fixed
count_birth_order4/5+ -0.1344 0.08117 -1.656 -0.2935 0.02468 fixed
count_birth_order5/5+ -0.08222 0.0746 -1.102 -0.2284 0.06399 fixed
count_birth_order5+/5+ -0.1868 0.05579 -3.349 -0.2961 -0.07746 fixed
sd_(Intercept).mother_pidlink 0.2954 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.887 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 15673 15746 -7826 15651 NA NA NA
12 15675 15755 -7825 15651 0.3398 1 0.5599
16 15678 15785 -7823 15646 4.718 4 0.3175
26 15692 15865 -7820 15640 6.553 10 0.7669

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Risk preference

Risk A

birthorder <- birthorder %>% mutate(outcome = riskA)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.147 0.1737 6.604 0.8068 1.488 fixed
poly(age, 3, raw = TRUE)1 -0.09252 0.01678 -5.513 -0.1254 -0.05963 fixed
poly(age, 3, raw = TRUE)2 0.002406 0.0004989 4.822 0.001428 0.003383 fixed
poly(age, 3, raw = TRUE)3 -0.00001908 0.000004653 -4.1 -0.00002819 -0.000009956 fixed
male -0.2332 0.01776 -13.13 -0.268 -0.1984 fixed
sibling_count3 0.003204 0.03507 0.09136 -0.06554 0.07195 fixed
sibling_count4 0.004187 0.03594 0.1165 -0.06626 0.07464 fixed
sibling_count5 -0.008956 0.03722 -0.2407 -0.0819 0.06399 fixed
sibling_count5+ 0.04878 0.02933 1.663 -0.008715 0.1063 fixed
sd_(Intercept).mother_pidlink 0.2501 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9575 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.149 0.1738 6.615 0.8089 1.49 fixed
birth_order 0.002926 0.003685 0.794 -0.004297 0.01015 fixed
poly(age, 3, raw = TRUE)1 -0.09344 0.01682 -5.555 -0.1264 -0.06047 fixed
poly(age, 3, raw = TRUE)2 0.002438 0.0005005 4.871 0.001457 0.003419 fixed
poly(age, 3, raw = TRUE)3 -0.00001937 0.000004668 -4.15 -0.00002852 -0.00001022 fixed
male -0.2333 0.01776 -13.14 -0.2681 -0.1985 fixed
sibling_count3 0.002364 0.03509 0.06739 -0.0664 0.07113 fixed
sibling_count4 0.002011 0.03604 0.05579 -0.06863 0.07265 fixed
sibling_count5 -0.01272 0.03751 -0.3391 -0.08624 0.0608 fixed
sibling_count5+ 0.03741 0.03264 1.146 -0.02656 0.1014 fixed
sd_(Intercept).mother_pidlink 0.2494 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9577 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.134 0.1743 6.508 0.7926 1.476 fixed
poly(age, 3, raw = TRUE)1 -0.09246 0.01683 -5.495 -0.1254 -0.05948 fixed
poly(age, 3, raw = TRUE)2 0.002401 0.0005007 4.795 0.001419 0.003382 fixed
poly(age, 3, raw = TRUE)3 -0.00001898 0.00000467 -4.064 -0.00002813 -0.000009823 fixed
male -0.2334 0.01776 -13.14 -0.2682 -0.1986 fixed
sibling_count3 0.00682 0.03566 0.1912 -0.06308 0.07672 fixed
sibling_count4 0.006205 0.03719 0.1669 -0.06668 0.07909 fixed
sibling_count5 -0.01895 0.03919 -0.4834 -0.09576 0.05787 fixed
sibling_count5+ 0.04327 0.03453 1.253 -0.02441 0.111 fixed
birth_order_nonlinear2 0.03422 0.02611 1.311 -0.01694 0.08539 fixed
birth_order_nonlinear3 -0.005286 0.0307 -0.1722 -0.06545 0.05488 fixed
birth_order_nonlinear4 0.015 0.03513 0.4268 -0.05386 0.08385 fixed
birth_order_nonlinear5 0.08311 0.03947 2.106 0.005747 0.1605 fixed
birth_order_nonlinear5+ 0.009131 0.03275 0.2789 -0.05505 0.07331 fixed
sd_(Intercept).mother_pidlink 0.2505 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9573 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.152 0.175 6.581 0.8086 1.495 fixed
poly(age, 3, raw = TRUE)1 -0.09299 0.01682 -5.528 -0.126 -0.06002 fixed
poly(age, 3, raw = TRUE)2 0.002411 0.0005005 4.817 0.00143 0.003392 fixed
poly(age, 3, raw = TRUE)3 -0.00001901 0.000004669 -4.072 -0.00002816 -0.000009861 fixed
male -0.2341 0.01775 -13.19 -0.2689 -0.1993 fixed
count_birth_order2/2 0.005908 0.05039 0.1173 -0.09284 0.1047 fixed
count_birth_order1/3 -0.04965 0.04725 -1.051 -0.1423 0.04296 fixed
count_birth_order2/3 0.02964 0.05332 0.5559 -0.07487 0.1342 fixed
count_birth_order3/3 0.08396 0.05874 1.429 -0.03117 0.1991 fixed
count_birth_order1/4 0.05371 0.05395 0.9955 -0.05203 0.1594 fixed
count_birth_order2/4 -0.01678 0.05658 -0.2965 -0.1277 0.09412 fixed
count_birth_order3/4 -0.09029 0.06134 -1.472 -0.2105 0.02993 fixed
count_birth_order4/4 0.07955 0.06589 1.207 -0.0496 0.2087 fixed
count_birth_order1/5 -0.02778 0.062 -0.4482 -0.1493 0.09372 fixed
count_birth_order2/5 0.04518 0.06404 0.7054 -0.08035 0.1707 fixed
count_birth_order3/5 0.01774 0.06616 0.2682 -0.1119 0.1474 fixed
count_birth_order4/5 -0.1898 0.06999 -2.712 -0.327 -0.05266 fixed
count_birth_order5/5 0.1188 0.07047 1.686 -0.01933 0.2569 fixed
count_birth_order1/5+ 0.02633 0.04975 0.5293 -0.07118 0.1238 fixed
count_birth_order2/5+ 0.1002 0.05139 1.949 -0.0005436 0.2009 fixed
count_birth_order3/5+ -0.006486 0.05003 -0.1296 -0.1045 0.09157 fixed
count_birth_order4/5+ 0.0823 0.049 1.68 -0.01373 0.1783 fixed
count_birth_order5/5+ 0.09391 0.04895 1.919 -0.00202 0.1898 fixed
count_birth_order5+/5+ 0.04251 0.03807 1.116 -0.03211 0.1171 fixed
sd_(Intercept).mother_pidlink 0.2503 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9567 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 34828 34909 -17403 34806 NA NA NA
12 34829 34918 -17403 34805 0.6324 1 0.4265
16 34831 34949 -17399 34799 6.352 4 0.1743
26 34825 35018 -17386 34773 25.71 10 0.004149

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.577 0.4493 3.51 0.6964 2.458 fixed
poly(age, 3, raw = TRUE)1 -0.134 0.05105 -2.626 -0.2341 -0.034 fixed
poly(age, 3, raw = TRUE)2 0.003493 0.001825 1.914 -0.00008355 0.007069 fixed
poly(age, 3, raw = TRUE)3 -0.00002836 0.0000207 -1.37 -0.00006894 0.00001222 fixed
male -0.2542 0.02616 -9.717 -0.3055 -0.2029 fixed
sibling_count3 -0.01637 0.04026 -0.4067 -0.09528 0.06253 fixed
sibling_count4 0.01001 0.04325 0.2314 -0.07475 0.09477 fixed
sibling_count5 0.02769 0.04899 0.5652 -0.06832 0.1237 fixed
sibling_count5+ 0.1299 0.04299 3.021 0.0456 0.2141 fixed
sd_(Intercept).mother_pidlink 0.2518 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9226 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.58 0.4494 3.516 0.6991 2.461 fixed
birth_order -0.003278 0.008588 -0.3817 -0.02011 0.01355 fixed
poly(age, 3, raw = TRUE)1 -0.1339 0.05105 -2.624 -0.234 -0.03387 fixed
poly(age, 3, raw = TRUE)2 0.003494 0.001825 1.915 -0.00008234 0.00707 fixed
poly(age, 3, raw = TRUE)3 -0.0000285 0.00002071 -1.376 -0.00006909 0.00001208 fixed
male -0.2541 0.02616 -9.71 -0.3054 -0.2028 fixed
sibling_count3 -0.01474 0.04049 -0.3641 -0.0941 0.06462 fixed
sibling_count4 0.01385 0.0444 0.3118 -0.07318 0.1009 fixed
sibling_count5 0.03389 0.05161 0.6566 -0.06727 0.1351 fixed
sibling_count5+ 0.1424 0.05406 2.634 0.03642 0.2483 fixed
sd_(Intercept).mother_pidlink 0.2515 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9228 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.557 0.4502 3.459 0.6751 2.44 fixed
poly(age, 3, raw = TRUE)1 -0.1325 0.05109 -2.593 -0.2326 -0.03235 fixed
poly(age, 3, raw = TRUE)2 0.003436 0.001826 1.881 -0.0001438 0.007016 fixed
poly(age, 3, raw = TRUE)3 -0.00002774 0.00002073 -1.338 -0.00006837 0.00001289 fixed
male -0.2537 0.02617 -9.693 -0.305 -0.2024 fixed
sibling_count3 -0.02883 0.04149 -0.6949 -0.1101 0.05249 fixed
sibling_count4 0.003924 0.04641 0.08454 -0.08704 0.09489 fixed
sibling_count5 0.01893 0.05466 0.3463 -0.08821 0.1261 fixed
sibling_count5+ 0.1267 0.05591 2.266 0.01712 0.2363 fixed
birth_order_nonlinear2 0.01925 0.03437 0.5601 -0.04811 0.0866 fixed
birth_order_nonlinear3 0.05381 0.04222 1.274 -0.02895 0.1366 fixed
birth_order_nonlinear4 -0.02119 0.05203 -0.4073 -0.1232 0.08078 fixed
birth_order_nonlinear5 0.03094 0.0651 0.4752 -0.09666 0.1585 fixed
birth_order_nonlinear5+ -0.00119 0.06463 -0.01842 -0.1279 0.1255 fixed
sd_(Intercept).mother_pidlink 0.2539 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9223 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.526 0.4507 3.385 0.6422 2.409 fixed
poly(age, 3, raw = TRUE)1 -0.1286 0.05114 -2.514 -0.2288 -0.02832 fixed
poly(age, 3, raw = TRUE)2 0.003327 0.001828 1.82 -0.0002567 0.00691 fixed
poly(age, 3, raw = TRUE)3 -0.00002686 0.00002075 -1.294 -0.00006754 0.00001381 fixed
male -0.2536 0.02619 -9.683 -0.3049 -0.2022 fixed
count_birth_order2/2 -0.008658 0.06139 -0.141 -0.129 0.1117 fixed
count_birth_order1/3 -0.1013 0.05393 -1.878 -0.207 0.00441 fixed
count_birth_order2/3 0.01465 0.05898 0.2484 -0.1009 0.1302 fixed
count_birth_order3/3 0.08765 0.06512 1.346 -0.03997 0.2153 fixed
count_birth_order1/4 0.05273 0.06664 0.7913 -0.07788 0.1834 fixed
count_birth_order2/4 -0.01304 0.06801 -0.1917 -0.1463 0.1203 fixed
count_birth_order3/4 -0.04041 0.07259 -0.5568 -0.1827 0.1019 fixed
count_birth_order4/4 0.02005 0.07383 0.2715 -0.1246 0.1647 fixed
count_birth_order1/5 0.1057 0.08996 1.175 -0.07063 0.282 fixed
count_birth_order2/5 -0.01872 0.09657 -0.1939 -0.208 0.1706 fixed
count_birth_order3/5 0.02719 0.0897 0.3031 -0.1486 0.203 fixed
count_birth_order4/5 0.0244 0.08554 0.2853 -0.1432 0.1921 fixed
count_birth_order5/5 -0.02271 0.08904 -0.2551 -0.1972 0.1518 fixed
count_birth_order1/5+ 0.1101 0.09201 1.197 -0.07023 0.2905 fixed
count_birth_order2/5+ 0.1728 0.08828 1.957 -0.0002416 0.3458 fixed
count_birth_order3/5+ 0.1931 0.08797 2.195 0.02068 0.3655 fixed
count_birth_order4/5+ -0.004091 0.08442 -0.04846 -0.1696 0.1614 fixed
count_birth_order5/5+ 0.1929 0.07866 2.452 0.03868 0.347 fixed
count_birth_order5+/5+ 0.1152 0.05806 1.985 0.001439 0.229 fixed
sd_(Intercept).mother_pidlink 0.2554 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9215 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 14660 14732 -7319 14638 NA NA NA
12 14661 14740 -7319 14637 0.1468 1 0.7016
16 14667 14772 -7317 14635 2.604 4 0.6262
26 14672 14844 -7310 14620 14.37 10 0.1569

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.549 0.4478 3.46 0.6717 2.427 fixed
poly(age, 3, raw = TRUE)1 -0.1301 0.05091 -2.555 -0.2298 -0.0303 fixed
poly(age, 3, raw = TRUE)2 0.003384 0.00182 1.859 -0.0001836 0.006951 fixed
poly(age, 3, raw = TRUE)3 -0.00002719 0.00002066 -1.316 -0.00006768 0.0000133 fixed
male -0.2574 0.02605 -9.882 -0.3085 -0.2064 fixed
sibling_count3 -0.03451 0.04357 -0.792 -0.1199 0.05089 fixed
sibling_count4 -0.03854 0.04574 -0.8425 -0.1282 0.05111 fixed
sibling_count5 0.02241 0.04858 0.4612 -0.07281 0.1176 fixed
sibling_count5+ 0.0725 0.04271 1.697 -0.01122 0.1562 fixed
sd_(Intercept).mother_pidlink 0.2535 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9221 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.545 0.4479 3.45 0.6672 2.423 fixed
birth_order 0.003578 0.007458 0.4798 -0.01104 0.01819 fixed
poly(age, 3, raw = TRUE)1 -0.1301 0.05091 -2.555 -0.2299 -0.03029 fixed
poly(age, 3, raw = TRUE)2 0.003378 0.00182 1.856 -0.00019 0.006945 fixed
poly(age, 3, raw = TRUE)3 -0.00002699 0.00002066 -1.306 -0.00006748 0.00001351 fixed
male -0.2575 0.02605 -9.886 -0.3086 -0.2065 fixed
sibling_count3 -0.03628 0.04373 -0.8296 -0.122 0.04943 fixed
sibling_count4 -0.04254 0.0465 -0.9149 -0.1337 0.0486 fixed
sibling_count5 0.01616 0.0503 0.3213 -0.08242 0.1147 fixed
sibling_count5+ 0.0592 0.05092 1.163 -0.04059 0.159 fixed
sd_(Intercept).mother_pidlink 0.2537 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9222 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.522 0.4487 3.392 0.6425 2.401 fixed
poly(age, 3, raw = TRUE)1 -0.128 0.05094 -2.513 -0.2279 -0.02816 fixed
poly(age, 3, raw = TRUE)2 0.003294 0.001822 1.808 -0.0002767 0.006864 fixed
poly(age, 3, raw = TRUE)3 -0.00002588 0.00002068 -1.252 -0.00006642 0.00001465 fixed
male -0.2576 0.02605 -9.888 -0.3086 -0.2065 fixed
sibling_count3 -0.05732 0.04465 -1.284 -0.1448 0.03021 fixed
sibling_count4 -0.06546 0.04836 -1.354 -0.1602 0.02932 fixed
sibling_count5 -0.0008842 0.05303 -0.01667 -0.1048 0.103 fixed
sibling_count5+ 0.0348 0.05268 0.6606 -0.06845 0.138 fixed
birth_order_nonlinear2 0.03458 0.03489 0.9911 -0.0338 0.103 fixed
birth_order_nonlinear3 0.09857 0.04209 2.342 0.01609 0.1811 fixed
birth_order_nonlinear4 0.0352 0.05032 0.6995 -0.06343 0.1338 fixed
birth_order_nonlinear5 0.003997 0.06168 0.06481 -0.1169 0.1249 fixed
birth_order_nonlinear5+ 0.07155 0.05752 1.244 -0.0412 0.1843 fixed
sd_(Intercept).mother_pidlink 0.2558 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9214 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.49 0.4492 3.317 0.6097 2.371 fixed
poly(age, 3, raw = TRUE)1 -0.1225 0.05098 -2.404 -0.2224 -0.02261 fixed
poly(age, 3, raw = TRUE)2 0.003108 0.001823 1.705 -0.0004651 0.006681 fixed
poly(age, 3, raw = TRUE)3 -0.00002392 0.0000207 -1.156 -0.00006449 0.00001665 fixed
male -0.2572 0.02606 -9.871 -0.3083 -0.2061 fixed
count_birth_order2/2 -0.01902 0.06687 -0.2844 -0.1501 0.112 fixed
count_birth_order1/3 -0.1438 0.05846 -2.46 -0.2584 -0.02923 fixed
count_birth_order2/3 0.003974 0.06319 0.0629 -0.1199 0.1278 fixed
count_birth_order3/3 0.08898 0.0708 1.257 -0.04978 0.2277 fixed
count_birth_order1/4 -0.01275 0.06936 -0.1838 -0.1487 0.1232 fixed
count_birth_order2/4 -0.1237 0.07019 -1.762 -0.2612 0.01391 fixed
count_birth_order3/4 -0.04247 0.07755 -0.5477 -0.1945 0.1095 fixed
count_birth_order4/4 0.01075 0.07811 0.1376 -0.1423 0.1638 fixed
count_birth_order1/5 0.01399 0.08174 0.1712 -0.1462 0.1742 fixed
count_birth_order2/5 0.009527 0.08838 0.1078 -0.1637 0.1827 fixed
count_birth_order3/5 0.1061 0.08479 1.251 -0.06012 0.2722 fixed
count_birth_order4/5 0.02947 0.08793 0.3351 -0.1429 0.2018 fixed
count_birth_order5/5 -0.0916 0.08745 -1.047 -0.263 0.07981 fixed
count_birth_order1/5+ 0.001461 0.07901 0.01849 -0.1534 0.1563 fixed
count_birth_order2/5+ 0.1332 0.08149 1.635 -0.02647 0.293 fixed
count_birth_order3/5+ 0.06555 0.08021 0.8172 -0.09166 0.2228 fixed
count_birth_order4/5+ -0.01491 0.07757 -0.1922 -0.1669 0.1371 fixed
count_birth_order5/5+ 0.07852 0.07906 0.9931 -0.07644 0.2335 fixed
count_birth_order5+/5+ 0.08841 0.05687 1.555 -0.02305 0.1999 fixed
sd_(Intercept).mother_pidlink 0.2571 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9206 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 14794 14866 -7386 14772 NA NA NA
12 14796 14875 -7386 14772 0.2304 1 0.6312
16 14797 14903 -7383 14765 6.297 4 0.178
26 14802 14973 -7375 14750 15.5 10 0.1149

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.476 0.455 3.243 0.5838 2.367 fixed
poly(age, 3, raw = TRUE)1 -0.1219 0.05172 -2.358 -0.2233 -0.02058 fixed
poly(age, 3, raw = TRUE)2 0.003027 0.00185 1.637 -0.000598 0.006652 fixed
poly(age, 3, raw = TRUE)3 -0.00002277 0.00002101 -1.084 -0.00006394 0.0000184 fixed
male -0.259 0.02642 -9.801 -0.3108 -0.2072 fixed
sibling_count3 -0.01595 0.03977 -0.4011 -0.09389 0.06199 fixed
sibling_count4 0.002577 0.04305 0.05986 -0.0818 0.08695 fixed
sibling_count5 0.06957 0.05029 1.383 -0.02899 0.1681 fixed
sibling_count5+ 0.1304 0.04352 2.997 0.04515 0.2157 fixed
sd_(Intercept).mother_pidlink 0.255 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.922 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.477 0.4551 3.246 0.5854 2.369 fixed
birth_order -0.002007 0.008824 -0.2275 -0.0193 0.01529 fixed
poly(age, 3, raw = TRUE)1 -0.1219 0.05173 -2.356 -0.2233 -0.02048 fixed
poly(age, 3, raw = TRUE)2 0.003028 0.00185 1.637 -0.0005978 0.006653 fixed
poly(age, 3, raw = TRUE)3 -0.00002285 0.00002101 -1.088 -0.00006403 0.00001833 fixed
male -0.2589 0.02643 -9.798 -0.3107 -0.2071 fixed
sibling_count3 -0.01493 0.04002 -0.3731 -0.09337 0.06351 fixed
sibling_count4 0.004905 0.04426 0.1108 -0.08183 0.09164 fixed
sibling_count5 0.0732 0.05277 1.387 -0.03022 0.1766 fixed
sibling_count5+ 0.1381 0.0549 2.515 0.03046 0.2456 fixed
sd_(Intercept).mother_pidlink 0.2549 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9221 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.46 0.4559 3.203 0.5665 2.354 fixed
poly(age, 3, raw = TRUE)1 -0.1207 0.05177 -2.332 -0.2222 -0.01925 fixed
poly(age, 3, raw = TRUE)2 0.002979 0.001851 1.609 -0.0006496 0.006607 fixed
poly(age, 3, raw = TRUE)3 -0.00002217 0.00002103 -1.054 -0.00006339 0.00001905 fixed
male -0.2584 0.02643 -9.775 -0.3102 -0.2066 fixed
sibling_count3 -0.02704 0.04106 -0.6587 -0.1075 0.05342 fixed
sibling_count4 -0.002276 0.04631 -0.04916 -0.09303 0.08848 fixed
sibling_count5 0.06283 0.0556 1.13 -0.04614 0.1718 fixed
sibling_count5+ 0.1202 0.05686 2.113 0.008714 0.2316 fixed
birth_order_nonlinear2 0.01531 0.03429 0.4465 -0.05189 0.08251 fixed
birth_order_nonlinear3 0.04702 0.04216 1.115 -0.03561 0.1297 fixed
birth_order_nonlinear4 -0.02296 0.05335 -0.4304 -0.1275 0.08161 fixed
birth_order_nonlinear5 0.02594 0.06801 0.3815 -0.1073 0.1592 fixed
birth_order_nonlinear5+ 0.01797 0.0667 0.2694 -0.1128 0.1487 fixed
sd_(Intercept).mother_pidlink 0.2571 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9217 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.418 0.4564 3.108 0.5238 2.313 fixed
poly(age, 3, raw = TRUE)1 -0.1144 0.05181 -2.207 -0.2159 -0.01281 fixed
poly(age, 3, raw = TRUE)2 0.002775 0.001853 1.497 -0.0008571 0.006406 fixed
poly(age, 3, raw = TRUE)3 -0.00002013 0.00002105 -0.9565 -0.00006139 0.00002112 fixed
male -0.2579 0.02644 -9.756 -0.3098 -0.2061 fixed
count_birth_order2/2 -0.03894 0.05986 -0.6506 -0.1563 0.07838 fixed
count_birth_order1/3 -0.1122 0.05335 -2.103 -0.2168 -0.007653 fixed
count_birth_order2/3 -0.005963 0.05896 -0.1011 -0.1215 0.1096 fixed
count_birth_order3/3 0.0912 0.06371 1.432 -0.03367 0.2161 fixed
count_birth_order1/4 0.0365 0.06697 0.5451 -0.09475 0.1678 fixed
count_birth_order2/4 0.00438 0.06821 0.06421 -0.1293 0.1381 fixed
count_birth_order3/4 -0.08139 0.07209 -1.129 -0.2227 0.05991 fixed
count_birth_order4/4 -0.01876 0.07401 -0.2534 -0.1638 0.1263 fixed
count_birth_order1/5 0.1228 0.09013 1.362 -0.0539 0.2994 fixed
count_birth_order2/5 0.01063 0.09959 0.1067 -0.1846 0.2058 fixed
count_birth_order3/5 0.05929 0.09337 0.6351 -0.1237 0.2423 fixed
count_birth_order4/5 0.09376 0.09014 1.04 -0.0829 0.2704 fixed
count_birth_order5/5 -0.02808 0.09563 -0.2937 -0.2155 0.1594 fixed
count_birth_order1/5+ 0.08835 0.09431 0.9368 -0.09649 0.2732 fixed
count_birth_order2/5+ 0.1599 0.09032 1.771 -0.0171 0.3369 fixed
count_birth_order3/5+ 0.1489 0.08977 1.659 -0.02704 0.3249 fixed
count_birth_order4/5+ -0.02845 0.0881 -0.3229 -0.2011 0.1442 fixed
count_birth_order5/5+ 0.1909 0.08075 2.365 0.03267 0.3492 fixed
count_birth_order5+/5+ 0.119 0.05923 2.01 0.002956 0.2351 fixed
sd_(Intercept).mother_pidlink 0.2591 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9205 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 14385 14457 -7181 14363 NA NA NA
12 14386 14465 -7181 14362 0.05235 1 0.819
16 14392 14497 -7180 14360 2.057 4 0.7253
26 14395 14565 -7171 14343 17.69 10 0.06035

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Risk B

birthorder <- birthorder %>% mutate(outcome = riskB)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.3885 0.1683 2.309 0.05867 0.7184 fixed
poly(age, 3, raw = TRUE)1 -0.02493 0.01621 -1.537 -0.05671 0.00685 fixed
poly(age, 3, raw = TRUE)2 0.0007458 0.0004805 1.552 -0.0001959 0.001688 fixed
poly(age, 3, raw = TRUE)3 -0.000007326 0.000004462 -1.642 -0.00001607 0.00000142 fixed
male -0.1892 0.01731 -10.93 -0.2231 -0.1553 fixed
sibling_count3 -0.02192 0.03395 -0.6458 -0.08846 0.04461 fixed
sibling_count4 -0.00809 0.03496 -0.2314 -0.07661 0.06043 fixed
sibling_count5 -0.03174 0.03599 -0.8819 -0.1023 0.0388 fixed
sibling_count5+ -0.04519 0.0285 -1.586 -0.1011 0.01067 fixed
sd_(Intercept).mother_pidlink 0.2164 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.969 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.3874 0.1683 2.301 0.05746 0.7173 fixed
birth_order -0.001201 0.003596 -0.3339 -0.008249 0.005847 fixed
poly(age, 3, raw = TRUE)1 -0.02453 0.01626 -1.509 -0.05639 0.007339 fixed
poly(age, 3, raw = TRUE)2 0.000732 0.0004823 1.518 -0.0002132 0.001677 fixed
poly(age, 3, raw = TRUE)3 -0.0000072 0.000004478 -1.608 -0.00001598 0.000001577 fixed
male -0.1892 0.01731 -10.93 -0.2231 -0.1552 fixed
sibling_count3 -0.02161 0.03396 -0.6363 -0.08817 0.04495 fixed
sibling_count4 -0.007222 0.03506 -0.206 -0.07594 0.06149 fixed
sibling_count5 -0.03024 0.03627 -0.8337 -0.1013 0.04085 fixed
sibling_count5+ -0.04059 0.03166 -1.282 -0.1026 0.02147 fixed
sd_(Intercept).mother_pidlink 0.2165 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.969 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.3921 0.1689 2.321 0.06105 0.7231 fixed
poly(age, 3, raw = TRUE)1 -0.02604 0.01627 -1.601 -0.05792 0.005844 fixed
poly(age, 3, raw = TRUE)2 0.0007753 0.0004825 1.607 -0.0001703 0.001721 fixed
poly(age, 3, raw = TRUE)3 -0.000007514 0.00000448 -1.677 -0.0000163 0.000001268 fixed
male -0.1895 0.01731 -10.95 -0.2234 -0.1555 fixed
sibling_count3 -0.03004 0.0345 -0.8706 -0.09766 0.03758 fixed
sibling_count4 -0.01974 0.03616 -0.546 -0.09061 0.05112 fixed
sibling_count5 -0.04581 0.0379 -1.209 -0.1201 0.02847 fixed
sibling_count5+ -0.06451 0.03346 -1.928 -0.1301 0.001068 fixed
birth_order_nonlinear2 0.01757 0.02529 0.6947 -0.032 0.06714 fixed
birth_order_nonlinear3 0.04294 0.02996 1.433 -0.01579 0.1017 fixed
birth_order_nonlinear4 0.02767 0.03418 0.8097 -0.03932 0.09466 fixed
birth_order_nonlinear5 0.0263 0.03854 0.6825 -0.04923 0.1018 fixed
birth_order_nonlinear5+ 0.03267 0.03179 1.028 -0.02964 0.09498 fixed
sd_(Intercept).mother_pidlink 0.217 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.969 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.4261 0.1695 2.513 0.09383 0.7585 fixed
poly(age, 3, raw = TRUE)1 -0.02658 0.01626 -1.635 -0.05845 0.005291 fixed
poly(age, 3, raw = TRUE)2 0.0007872 0.0004824 1.632 -0.0001582 0.001733 fixed
poly(age, 3, raw = TRUE)3 -0.00000757 0.00000448 -1.69 -0.00001635 0.000001211 fixed
male -0.1896 0.0173 -10.96 -0.2235 -0.1556 fixed
count_birth_order2/2 -0.05705 0.04918 -1.16 -0.1534 0.03933 fixed
count_birth_order1/3 -0.1249 0.04595 -2.718 -0.215 -0.03484 fixed
count_birth_order2/3 0.0004879 0.05123 0.009525 -0.09991 0.1009 fixed
count_birth_order3/3 0.06406 0.05765 1.111 -0.04893 0.1771 fixed
count_birth_order1/4 0.03162 0.05249 0.6024 -0.07126 0.1345 fixed
count_birth_order2/4 -0.04005 0.05544 -0.7225 -0.1487 0.0686 fixed
count_birth_order3/4 -0.1284 0.06086 -2.109 -0.2476 -0.009089 fixed
count_birth_order4/4 -0.001447 0.06385 -0.02266 -0.1266 0.1237 fixed
count_birth_order1/5 -0.1393 0.05994 -2.324 -0.2568 -0.02183 fixed
count_birth_order2/5 -0.06536 0.06223 -1.05 -0.1873 0.05661 fixed
count_birth_order3/5 0.00197 0.06385 0.03086 -0.1232 0.1271 fixed
count_birth_order4/5 -0.05504 0.06836 -0.8053 -0.189 0.07893 fixed
count_birth_order5/5 0.03196 0.06921 0.4618 -0.1037 0.1676 fixed
count_birth_order1/5+ -0.08178 0.04827 -1.694 -0.1764 0.01282 fixed
count_birth_order2/5+ -0.05255 0.04983 -1.055 -0.1502 0.04512 fixed
count_birth_order3/5+ -0.04667 0.04888 -0.9549 -0.1425 0.04912 fixed
count_birth_order4/5+ -0.06846 0.04795 -1.428 -0.1624 0.02552 fixed
count_birth_order5/5+ -0.09239 0.04793 -1.928 -0.1863 0.00154 fixed
count_birth_order5+/5+ -0.05918 0.03721 -1.591 -0.1321 0.01374 fixed
sd_(Intercept).mother_pidlink 0.2169 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9685 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 37149 37232 -18564 37127 NA NA NA
12 37151 37241 -18564 37127 0.1115 1 0.7384
16 37157 37277 -18562 37125 2.188 4 0.7013
26 37153 37348 -18550 37101 23.98 10 0.007645

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.049 0.4283 2.45 0.2097 1.888 fixed
poly(age, 3, raw = TRUE)1 -0.0891 0.04851 -1.837 -0.1842 0.005985 fixed
poly(age, 3, raw = TRUE)2 0.002848 0.001728 1.648 -0.0005399 0.006235 fixed
poly(age, 3, raw = TRUE)3 -0.0000287 0.00001954 -1.469 -0.00006699 0.000009593 fixed
male -0.2247 0.02515 -8.938 -0.274 -0.1755 fixed
sibling_count3 -0.07825 0.03816 -2.051 -0.153 -0.003456 fixed
sibling_count4 -0.02742 0.04083 -0.6716 -0.1074 0.0526 fixed
sibling_count5 -0.05557 0.04624 -1.202 -0.1462 0.03506 fixed
sibling_count5+ -0.01853 0.04043 -0.4583 -0.09776 0.06071 fixed
sd_(Intercept).mother_pidlink 0.1238 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9313 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.052 0.4284 2.455 0.2119 1.891 fixed
birth_order -0.00229 0.008256 -0.2774 -0.01847 0.01389 fixed
poly(age, 3, raw = TRUE)1 -0.08908 0.04852 -1.836 -0.1842 0.006012 fixed
poly(age, 3, raw = TRUE)2 0.002851 0.001729 1.649 -0.0005368 0.006239 fixed
poly(age, 3, raw = TRUE)3 -0.00002882 0.00001954 -1.475 -0.00006713 0.000009482 fixed
male -0.2247 0.02515 -8.934 -0.274 -0.1754 fixed
sibling_count3 -0.07714 0.03837 -2.011 -0.1523 -0.001941 fixed
sibling_count4 -0.02485 0.04188 -0.5933 -0.1069 0.05723 fixed
sibling_count5 -0.05135 0.04868 -1.055 -0.1468 0.04407 fixed
sibling_count5+ -0.01003 0.05071 -0.1979 -0.1094 0.08936 fixed
sd_(Intercept).mother_pidlink 0.1234 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9315 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.026 0.429 2.391 0.1848 1.866 fixed
poly(age, 3, raw = TRUE)1 -0.08819 0.04855 -1.816 -0.1833 0.006968 fixed
poly(age, 3, raw = TRUE)2 0.002811 0.00173 1.625 -0.0005792 0.006202 fixed
poly(age, 3, raw = TRUE)3 -0.00002829 0.00001956 -1.446 -0.00006663 0.00001005 fixed
male -0.2244 0.02515 -8.92 -0.2737 -0.1751 fixed
sibling_count3 -0.08624 0.03928 -2.196 -0.1632 -0.009255 fixed
sibling_count4 -0.03939 0.04378 -0.8997 -0.1252 0.04641 fixed
sibling_count5 -0.05825 0.05154 -1.13 -0.1593 0.04278 fixed
sibling_count5+ -0.0163 0.0524 -0.311 -0.119 0.08641 fixed
birth_order_nonlinear2 0.04954 0.03309 1.497 -0.01531 0.1144 fixed
birth_order_nonlinear3 0.03766 0.04077 0.9236 -0.04225 0.1176 fixed
birth_order_nonlinear4 0.03502 0.05007 0.6993 -0.06313 0.1332 fixed
birth_order_nonlinear5 -0.02212 0.06288 -0.3517 -0.1454 0.1011 fixed
birth_order_nonlinear5+ 0.01008 0.06199 0.1626 -0.1114 0.1316 fixed
sd_(Intercept).mother_pidlink 0.1225 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9316 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.9659 0.4297 2.248 0.1237 1.808 fixed
poly(age, 3, raw = TRUE)1 -0.08106 0.04864 -1.667 -0.1764 0.01426 fixed
poly(age, 3, raw = TRUE)2 0.002578 0.001733 1.487 -0.000819 0.005975 fixed
poly(age, 3, raw = TRUE)3 -0.00002589 0.0000196 -1.321 -0.00006431 0.00001253 fixed
male -0.2227 0.02516 -8.853 -0.2721 -0.1734 fixed
count_birth_order2/2 0.02637 0.05933 0.4445 -0.08992 0.1427 fixed
count_birth_order1/3 -0.1491 0.05167 -2.885 -0.2503 -0.04779 fixed
count_birth_order2/3 -0.006833 0.05633 -0.1213 -0.1172 0.1036 fixed
count_birth_order3/3 -0.005061 0.06322 -0.08006 -0.129 0.1189 fixed
count_birth_order1/4 0.007126 0.06334 0.1125 -0.117 0.1313 fixed
count_birth_order2/4 -0.03101 0.06571 -0.4719 -0.1598 0.09778 fixed
count_birth_order3/4 -0.07549 0.06934 -1.089 -0.2114 0.06041 fixed
count_birth_order4/4 0.02371 0.07167 0.3308 -0.1168 0.1642 fixed
count_birth_order1/5 0.0343 0.08707 0.3939 -0.1363 0.2049 fixed
count_birth_order2/5 -0.1382 0.0915 -1.511 -0.3176 0.0411 fixed
count_birth_order3/5 0.06142 0.08583 0.7156 -0.1068 0.2297 fixed
count_birth_order4/5 -0.08468 0.08257 -1.025 -0.2465 0.07716 fixed
count_birth_order5/5 -0.1152 0.08712 -1.322 -0.2859 0.05555 fixed
count_birth_order1/5+ -0.05294 0.08612 -0.6147 -0.2217 0.1159 fixed
count_birth_order2/5+ 0.1103 0.08418 1.31 -0.0547 0.2753 fixed
count_birth_order3/5+ -0.07246 0.08492 -0.8532 -0.2389 0.09399 fixed
count_birth_order4/5+ 0.01143 0.07976 0.1433 -0.1449 0.1678 fixed
count_birth_order5/5+ -0.02858 0.07532 -0.3794 -0.1762 0.119 fixed
count_birth_order5+/5+ -0.01489 0.05589 -0.2664 -0.1244 0.09465 fixed
sd_(Intercept).mother_pidlink 0.1247 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9309 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 15219 15292 -7598 15197 NA NA NA
12 15221 15300 -7598 15197 0.07815 1 0.7798
16 15226 15332 -7597 15194 3.215 4 0.5226
26 15230 15403 -7589 15178 15.45 10 0.1164

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.028 0.4272 2.407 0.1909 1.866 fixed
poly(age, 3, raw = TRUE)1 -0.08594 0.04842 -1.775 -0.1808 0.008949 fixed
poly(age, 3, raw = TRUE)2 0.00276 0.001725 1.6 -0.0006219 0.006141 fixed
poly(age, 3, raw = TRUE)3 -0.00002778 0.00001951 -1.424 -0.00006601 0.00001045 fixed
male -0.2282 0.02506 -9.107 -0.2773 -0.1791 fixed
sibling_count3 -0.09326 0.04133 -2.256 -0.1743 -0.01225 fixed
sibling_count4 -0.05237 0.04336 -1.208 -0.1374 0.03261 fixed
sibling_count5 -0.05936 0.04608 -1.288 -0.1497 0.03096 fixed
sibling_count5+ -0.04957 0.04042 -1.226 -0.1288 0.02965 fixed
sd_(Intercept).mother_pidlink 0.1245 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9322 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.03 0.4274 2.41 0.1922 1.867 fixed
birth_order -0.00116 0.007178 -0.1616 -0.01523 0.01291 fixed
poly(age, 3, raw = TRUE)1 -0.08597 0.04842 -1.775 -0.1809 0.008932 fixed
poly(age, 3, raw = TRUE)2 0.002763 0.001726 1.601 -0.0006194 0.006145 fixed
poly(age, 3, raw = TRUE)3 -0.00002786 0.00001951 -1.428 -0.0000661 0.00001039 fixed
male -0.2282 0.02506 -9.105 -0.2773 -0.1791 fixed
sibling_count3 -0.0927 0.04148 -2.235 -0.174 -0.01141 fixed
sibling_count4 -0.05113 0.04404 -1.161 -0.1374 0.03518 fixed
sibling_count5 -0.0574 0.04767 -1.204 -0.1508 0.03603 fixed
sibling_count5+ -0.04536 0.04809 -0.9432 -0.1396 0.0489 fixed
sd_(Intercept).mother_pidlink 0.1245 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9322 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.002 0.4279 2.342 0.1635 1.841 fixed
poly(age, 3, raw = TRUE)1 -0.08492 0.04844 -1.753 -0.1799 0.01002 fixed
poly(age, 3, raw = TRUE)2 0.002714 0.001726 1.572 -0.0006696 0.006098 fixed
poly(age, 3, raw = TRUE)3 -0.00002717 0.00001952 -1.391 -0.00006544 0.0000111 fixed
male -0.228 0.02506 -9.096 -0.2771 -0.1789 fixed
sibling_count3 -0.1039 0.04233 -2.456 -0.1869 -0.02098 fixed
sibling_count4 -0.06911 0.04576 -1.51 -0.1588 0.02058 fixed
sibling_count5 -0.0629 0.05026 -1.251 -0.1614 0.03561 fixed
sibling_count5+ -0.05826 0.04978 -1.17 -0.1558 0.03931 fixed
birth_order_nonlinear2 0.05573 0.03362 1.657 -0.01017 0.1216 fixed
birth_order_nonlinear3 0.04984 0.04063 1.227 -0.0298 0.1295 fixed
birth_order_nonlinear4 0.04656 0.04858 0.9584 -0.04865 0.1418 fixed
birth_order_nonlinear5 -0.03878 0.05997 -0.6466 -0.1563 0.07877 fixed
birth_order_nonlinear5+ 0.0368 0.05516 0.667 -0.07132 0.1449 fixed
sd_(Intercept).mother_pidlink 0.1244 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9322 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.9712 0.4283 2.267 0.1317 1.811 fixed
poly(age, 3, raw = TRUE)1 -0.0814 0.04848 -1.679 -0.1764 0.01361 fixed
poly(age, 3, raw = TRUE)2 0.002602 0.001728 1.506 -0.0007843 0.005989 fixed
poly(age, 3, raw = TRUE)3 -0.00002605 0.00001954 -1.333 -0.00006436 0.00001226 fixed
male -0.2291 0.02507 -9.14 -0.2782 -0.18 fixed
count_birth_order2/2 0.04849 0.06493 0.7467 -0.07878 0.1758 fixed
count_birth_order1/3 -0.1679 0.05589 -3.005 -0.2775 -0.05838 fixed
count_birth_order2/3 -0.02805 0.06057 -0.4631 -0.1468 0.09067 fixed
count_birth_order3/3 0.0296 0.06842 0.4325 -0.1045 0.1637 fixed
count_birth_order1/4 -0.005126 0.06627 -0.07735 -0.135 0.1248 fixed
count_birth_order2/4 -0.04986 0.06714 -0.7426 -0.1814 0.08173 fixed
count_birth_order3/4 -0.1197 0.07463 -1.605 -0.266 0.02652 fixed
count_birth_order4/4 0.02688 0.07657 0.3511 -0.1232 0.177 fixed
count_birth_order1/5 -0.0005703 0.07891 -0.007228 -0.1552 0.1541 fixed
count_birth_order2/5 -0.06387 0.08439 -0.7568 -0.2293 0.1015 fixed
count_birth_order3/5 0.05279 0.08275 0.6379 -0.1094 0.215 fixed
count_birth_order4/5 -0.1065 0.08431 -1.264 -0.2718 0.0587 fixed
count_birth_order5/5 -0.1147 0.08589 -1.335 -0.283 0.05365 fixed
count_birth_order1/5+ -0.06879 0.07534 -0.9131 -0.2164 0.07887 fixed
count_birth_order2/5+ 0.04887 0.07825 0.6245 -0.1045 0.2022 fixed
count_birth_order3/5+ -0.07751 0.07663 -1.011 -0.2277 0.07269 fixed
count_birth_order4/5+ -0.0007247 0.07385 -0.009813 -0.1455 0.144 fixed
count_birth_order5/5+ -0.09299 0.07634 -1.218 -0.2426 0.05664 fixed
count_birth_order5+/5+ -0.02458 0.0547 -0.4494 -0.1318 0.08263 fixed
sd_(Intercept).mother_pidlink 0.1213 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9321 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 15371 15444 -7675 15349 NA NA NA
12 15373 15453 -7675 15349 0.02676 1 0.8701
16 15376 15482 -7672 15344 5.218 4 0.2657
26 15381 15553 -7664 15329 15.32 10 0.1209

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.9132 0.4324 2.112 0.06576 1.761 fixed
poly(age, 3, raw = TRUE)1 -0.07301 0.049 -1.49 -0.1691 0.02304 fixed
poly(age, 3, raw = TRUE)2 0.002181 0.001747 1.249 -0.001243 0.005604 fixed
poly(age, 3, raw = TRUE)3 -0.00002022 0.00001976 -1.023 -0.00005894 0.0000185 fixed
male -0.2238 0.02536 -8.828 -0.2735 -0.1741 fixed
sibling_count3 -0.04656 0.03761 -1.238 -0.1203 0.02716 fixed
sibling_count4 -0.01363 0.04054 -0.3363 -0.09309 0.06582 fixed
sibling_count5 0.001808 0.04729 0.03824 -0.09087 0.09449 fixed
sibling_count5+ -0.01265 0.04078 -0.3103 -0.09257 0.06727 fixed
sd_(Intercept).mother_pidlink 0.1348 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9277 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.9127 0.4325 2.111 0.06512 1.76 fixed
birth_order 0.000456 0.008519 0.05353 -0.01624 0.01715 fixed
poly(age, 3, raw = TRUE)1 -0.07302 0.04901 -1.49 -0.1691 0.02304 fixed
poly(age, 3, raw = TRUE)2 0.00218 0.001747 1.248 -0.001243 0.005604 fixed
poly(age, 3, raw = TRUE)3 -0.00002019 0.00001976 -1.022 -0.00005893 0.00001854 fixed
male -0.2238 0.02536 -8.827 -0.2735 -0.1741 fixed
sibling_count3 -0.04678 0.03784 -1.236 -0.1209 0.02738 fixed
sibling_count4 -0.01414 0.04163 -0.3396 -0.09574 0.06746 fixed
sibling_count5 0.001006 0.04962 0.02027 -0.09624 0.09825 fixed
sibling_count5+ -0.01433 0.05143 -0.2786 -0.1151 0.08647 fixed
sd_(Intercept).mother_pidlink 0.1351 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9278 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.8882 0.433 2.051 0.03954 1.737 fixed
poly(age, 3, raw = TRUE)1 -0.07193 0.04904 -1.467 -0.168 0.02418 fixed
poly(age, 3, raw = TRUE)2 0.002137 0.001748 1.222 -0.001289 0.005562 fixed
poly(age, 3, raw = TRUE)3 -0.00001967 0.00001978 -0.9946 -0.00005843 0.00001909 fixed
male -0.2233 0.02537 -8.805 -0.2731 -0.1736 fixed
sibling_count3 -0.0543 0.03876 -1.401 -0.1303 0.02167 fixed
sibling_count4 -0.02342 0.04356 -0.5376 -0.1088 0.06196 fixed
sibling_count5 -0.001756 0.05226 -0.0336 -0.1042 0.1007 fixed
sibling_count5+ -0.016 0.05321 -0.3006 -0.1203 0.0883 fixed
birth_order_nonlinear2 0.04845 0.03294 1.471 -0.01611 0.113 fixed
birth_order_nonlinear3 0.03917 0.04067 0.9632 -0.04053 0.1189 fixed
birth_order_nonlinear4 0.02205 0.05151 0.428 -0.07892 0.123 fixed
birth_order_nonlinear5 -0.004136 0.06549 -0.06317 -0.1325 0.1242 fixed
birth_order_nonlinear5+ 0.02085 0.06395 0.3261 -0.1045 0.1462 fixed
sd_(Intercept).mother_pidlink 0.1348 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9279 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.831 0.4338 1.915 -0.01936 1.681 fixed
poly(age, 3, raw = TRUE)1 -0.06428 0.04913 -1.308 -0.1606 0.03203 fixed
poly(age, 3, raw = TRUE)2 0.001883 0.001752 1.075 -0.00155 0.005316 fixed
poly(age, 3, raw = TRUE)3 -0.00001703 0.00001982 -0.8592 -0.00005588 0.00002182 fixed
male -0.2212 0.02538 -8.718 -0.271 -0.1715 fixed
count_birth_order2/2 0.005957 0.0577 0.1032 -0.1071 0.119 fixed
count_birth_order1/3 -0.1323 0.0509 -2.599 -0.232 -0.03252 fixed
count_birth_order2/3 0.01953 0.0563 0.3468 -0.09081 0.1299 fixed
count_birth_order3/3 0.03569 0.06196 0.576 -0.08575 0.1571 fixed
count_birth_order1/4 0.02198 0.06356 0.3458 -0.1026 0.1465 fixed
count_birth_order2/4 -0.01268 0.06554 -0.1935 -0.1411 0.1158 fixed
count_birth_order3/4 -0.0643 0.06839 -0.9402 -0.1984 0.06975 fixed
count_birth_order4/4 0.001813 0.07231 0.02506 -0.1399 0.1435 fixed
count_birth_order1/5 0.08837 0.08671 1.019 -0.08157 0.2583 fixed
count_birth_order2/5 -0.07422 0.0942 -0.7878 -0.2588 0.1104 fixed
count_birth_order3/5 0.06378 0.08976 0.7106 -0.1121 0.2397 fixed
count_birth_order4/5 -0.01829 0.08662 -0.2112 -0.1881 0.1515 fixed
count_birth_order5/5 -0.06024 0.09297 -0.6479 -0.2425 0.122 fixed
count_birth_order1/5+ -0.07076 0.08852 -0.7993 -0.2443 0.1027 fixed
count_birth_order2/5+ 0.09919 0.08581 1.156 -0.069 0.2674 fixed
count_birth_order3/5+ -0.06262 0.08586 -0.7294 -0.2309 0.1056 fixed
count_birth_order4/5+ -0.01431 0.08371 -0.1709 -0.1784 0.1498 fixed
count_birth_order5/5+ -0.01093 0.07712 -0.1418 -0.1621 0.1402 fixed
count_birth_order5+/5+ -0.01065 0.05681 -0.1875 -0.122 0.1007 fixed
sd_(Intercept).mother_pidlink 0.1365 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9273 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 14888 14961 -7433 14866 NA NA NA
12 14890 14970 -7433 14866 0.002696 1 0.9586
16 14896 15001 -7432 14864 2.661 4 0.6161
26 14901 15073 -7424 14849 14.83 10 0.1385

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Educational Attainment

Years of Education - z-standardized

birthorder <- birthorder %>% mutate(outcome = years_of_education_z)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.688 0.1288 -20.87 -2.941 -2.436 fixed
poly(age, 3, raw = TRUE)1 0.2633 0.01224 21.51 0.2393 0.2873 fixed
poly(age, 3, raw = TRUE)2 -0.006974 0.0003591 -19.42 -0.007678 -0.00627 fixed
poly(age, 3, raw = TRUE)3 0.00005177 0.000003294 15.71 0.00004531 0.00005823 fixed
male -0.02813 0.013 -2.163 -0.05362 -0.002644 fixed
sibling_count3 0.07051 0.03368 2.093 0.004496 0.1365 fixed
sibling_count4 0.0338 0.03529 0.9576 -0.03538 0.103 fixed
sibling_count5 0.02754 0.0374 0.7364 -0.04577 0.1008 fixed
sibling_count5+ -0.2199 0.02876 -7.648 -0.2763 -0.1636 fixed
sd_(Intercept).mother_pidlink 0.6782 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.6413 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.696 0.1288 -20.93 -2.949 -2.444 fixed
birth_order 0.006501 0.002997 2.169 0.0006277 0.01237 fixed
poly(age, 3, raw = TRUE)1 0.2614 0.01227 21.3 0.2373 0.2854 fixed
poly(age, 3, raw = TRUE)2 -0.006884 0.0003615 -19.04 -0.007592 -0.006175 fixed
poly(age, 3, raw = TRUE)3 0.00005089 0.00000332 15.33 0.00004438 0.00005739 fixed
male -0.0283 0.013 -2.177 -0.05378 -0.002824 fixed
sibling_count3 0.07026 0.0337 2.085 0.004219 0.1363 fixed
sibling_count4 0.0308 0.03534 0.8716 -0.03846 0.1001 fixed
sibling_count5 0.02083 0.03755 0.5546 -0.05277 0.09442 fixed
sibling_count5+ -0.2435 0.03076 -7.917 -0.3038 -0.1832 fixed
sd_(Intercept).mother_pidlink 0.6792 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.6407 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.665 0.1291 -20.64 -2.918 -2.411 fixed
poly(age, 3, raw = TRUE)1 0.2613 0.01227 21.31 0.2373 0.2854 fixed
poly(age, 3, raw = TRUE)2 -0.006871 0.0003613 -19.02 -0.007579 -0.006163 fixed
poly(age, 3, raw = TRUE)3 0.00005046 0.000003321 15.2 0.00004396 0.00005697 fixed
male -0.02841 0.01299 -2.188 -0.05386 -0.002961 fixed
sibling_count3 0.08287 0.03392 2.443 0.01639 0.1494 fixed
sibling_count4 0.05735 0.0358 1.602 -0.01282 0.1275 fixed
sibling_count5 0.05175 0.03825 1.353 -0.02322 0.1267 fixed
sibling_count5+ -0.2166 0.03162 -6.85 -0.2786 -0.1546 fixed
birth_order_nonlinear2 -0.04969 0.01841 -2.698 -0.08578 -0.0136 fixed
birth_order_nonlinear3 -0.07061 0.02142 -3.296 -0.1126 -0.02862 fixed
birth_order_nonlinear4 -0.08076 0.02452 -3.293 -0.1288 -0.03269 fixed
birth_order_nonlinear5 -0.02921 0.02787 -1.048 -0.08383 0.02541 fixed
birth_order_nonlinear5+ 0.0102 0.02471 0.4128 -0.03823 0.05863 fixed
sd_(Intercept).mother_pidlink 0.6799 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.6398 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.655 0.1297 -20.47 -2.909 -2.401 fixed
poly(age, 3, raw = TRUE)1 0.261 0.01227 21.27 0.237 0.2851 fixed
poly(age, 3, raw = TRUE)2 -0.006856 0.0003615 -18.96 -0.007565 -0.006147 fixed
poly(age, 3, raw = TRUE)3 0.00005029 0.000003325 15.12 0.00004377 0.00005681 fixed
male -0.02811 0.01299 -2.165 -0.05357 -0.002659 fixed
count_birth_order2/2 -0.07379 0.03615 -2.041 -0.1446 -0.002938 fixed
count_birth_order1/3 0.07773 0.03991 1.948 -0.0004907 0.156 fixed
count_birth_order2/3 0.01959 0.04385 0.4469 -0.06635 0.1055 fixed
count_birth_order3/3 0.004158 0.04807 0.08651 -0.09005 0.09837 fixed
count_birth_order1/4 0.03963 0.04447 0.8911 -0.04753 0.1268 fixed
count_birth_order2/4 -0.01847 0.04706 -0.3925 -0.1107 0.07377 fixed
count_birth_order3/4 -0.003081 0.0501 -0.0615 -0.1013 0.09511 fixed
count_birth_order4/4 -0.006753 0.05219 -0.1294 -0.109 0.09554 fixed
count_birth_order1/5 0.03106 0.05016 0.6191 -0.06726 0.1294 fixed
count_birth_order2/5 0.004568 0.05248 0.08705 -0.09828 0.1074 fixed
count_birth_order3/5 -0.09389 0.05304 -1.77 -0.1978 0.01008 fixed
count_birth_order4/5 -0.04342 0.05613 -0.7736 -0.1534 0.06658 fixed
count_birth_order5/5 0.1147 0.0571 2.008 0.00277 0.2266 fixed
count_birth_order1/5+ -0.2286 0.0397 -5.758 -0.3064 -0.1508 fixed
count_birth_order2/5+ -0.2492 0.04077 -6.112 -0.3291 -0.1693 fixed
count_birth_order3/5+ -0.2759 0.03996 -6.904 -0.3542 -0.1976 fixed
count_birth_order4/5+ -0.3126 0.03947 -7.919 -0.3899 -0.2352 fixed
count_birth_order5/5+ -0.2873 0.03954 -7.266 -0.3648 -0.2098 fixed
count_birth_order5+/5+ -0.2149 0.03384 -6.349 -0.2812 -0.1485 fixed
sd_(Intercept).mother_pidlink 0.6799 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.6397 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 36298 36381 -18138 36276 NA NA NA
12 36295 36386 -18136 36271 4.685 1 0.03043
16 36281 36402 -18125 36249 22.33 4 0.0001724
26 36288 36485 -18118 36236 13.46 10 0.1991

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.736 0.309 -21.8 -7.342 -6.13 fixed
poly(age, 3, raw = TRUE)1 0.7187 0.03506 20.5 0.65 0.7874 fixed
poly(age, 3, raw = TRUE)2 -0.02238 0.001252 -17.87 -0.02483 -0.01993 fixed
poly(age, 3, raw = TRUE)3 0.0002241 0.0000142 15.78 0.0001963 0.000252 fixed
male -0.07737 0.0178 -4.347 -0.1123 -0.04248 fixed
sibling_count3 0.01365 0.03226 0.423 -0.04959 0.07688 fixed
sibling_count4 -0.08132 0.03529 -2.304 -0.1505 -0.01215 fixed
sibling_count5 -0.1649 0.04105 -4.017 -0.2454 -0.08447 fixed
sibling_count5+ -0.3901 0.03602 -10.83 -0.4607 -0.3195 fixed
sd_(Intercept).mother_pidlink 0.5155 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.5677 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.749 0.3088 -21.86 -7.354 -6.144 fixed
birth_order 0.01624 0.006102 2.662 0.004285 0.0282 fixed
poly(age, 3, raw = TRUE)1 0.7172 0.03504 20.47 0.6485 0.7858 fixed
poly(age, 3, raw = TRUE)2 -0.02234 0.001251 -17.85 -0.02479 -0.01988 fixed
poly(age, 3, raw = TRUE)3 0.0002243 0.00001419 15.81 0.0001965 0.0002522 fixed
male -0.07822 0.01779 -4.398 -0.1131 -0.04336 fixed
sibling_count3 0.004941 0.03243 0.1524 -0.05863 0.06851 fixed
sibling_count4 -0.1021 0.03616 -2.824 -0.173 -0.03126 fixed
sibling_count5 -0.198 0.04291 -4.615 -0.2821 -0.1139 fixed
sibling_count5+ -0.4559 0.04371 -10.43 -0.5415 -0.3702 fixed
sd_(Intercept).mother_pidlink 0.5166 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.5667 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.665 0.3096 -21.53 -7.271 -6.058 fixed
poly(age, 3, raw = TRUE)1 0.712 0.03506 20.31 0.6433 0.7808 fixed
poly(age, 3, raw = TRUE)2 -0.02217 0.001252 -17.71 -0.02462 -0.01971 fixed
poly(age, 3, raw = TRUE)3 0.0002228 0.00001419 15.69 0.0001949 0.0002506 fixed
male -0.07857 0.01776 -4.425 -0.1134 -0.04377 fixed
sibling_count3 0.005269 0.03283 0.1605 -0.05908 0.06962 fixed
sibling_count4 -0.1132 0.03708 -3.054 -0.1859 -0.04056 fixed
sibling_count5 -0.2355 0.04429 -5.317 -0.3223 -0.1487 fixed
sibling_count5+ -0.4792 0.04459 -10.74 -0.5666 -0.3918 fixed
birth_order_nonlinear2 -0.05353 0.02201 -2.433 -0.09667 -0.0104 fixed
birth_order_nonlinear3 0.02796 0.02736 1.022 -0.02568 0.08159 fixed
birth_order_nonlinear4 0.0843 0.03424 2.462 0.01719 0.1514 fixed
birth_order_nonlinear5 0.1608 0.04246 3.786 0.07755 0.244 fixed
birth_order_nonlinear5+ 0.07327 0.04454 1.645 -0.01403 0.1606 fixed
sd_(Intercept).mother_pidlink 0.5157 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.5657 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.676 0.3105 -21.5 -7.285 -6.067 fixed
poly(age, 3, raw = TRUE)1 0.713 0.03516 20.28 0.6441 0.7819 fixed
poly(age, 3, raw = TRUE)2 -0.02222 0.001256 -17.69 -0.02468 -0.01976 fixed
poly(age, 3, raw = TRUE)3 0.0002235 0.00001424 15.69 0.0001955 0.0002514 fixed
male -0.07799 0.01778 -4.386 -0.1128 -0.04314 fixed
count_birth_order2/2 -0.03586 0.04032 -0.8893 -0.1149 0.04317 fixed
count_birth_order1/3 0.03154 0.03934 0.8017 -0.04557 0.1087 fixed
count_birth_order2/3 -0.06423 0.04233 -1.517 -0.1472 0.01873 fixed
count_birth_order3/3 0.02724 0.04677 0.5824 -0.06443 0.1189 fixed
count_birth_order1/4 -0.1151 0.0475 -2.423 -0.2082 -0.02198 fixed
count_birth_order2/4 -0.1695 0.04852 -3.495 -0.2646 -0.07446 fixed
count_birth_order3/4 -0.0516 0.05091 -1.014 -0.1514 0.04818 fixed
count_birth_order4/4 -0.03001 0.05286 -0.5676 -0.1336 0.0736 fixed
count_birth_order1/5 -0.2277 0.06389 -3.564 -0.3529 -0.1025 fixed
count_birth_order2/5 -0.2326 0.06767 -3.437 -0.3652 -0.09994 fixed
count_birth_order3/5 -0.2085 0.06323 -3.297 -0.3324 -0.08455 fixed
count_birth_order4/5 -0.1802 0.06188 -2.912 -0.3015 -0.05892 fixed
count_birth_order5/5 -0.06835 0.06392 -1.069 -0.1936 0.05694 fixed
count_birth_order1/5+ -0.5118 0.06239 -8.204 -0.6341 -0.3895 fixed
count_birth_order2/5+ -0.5145 0.06185 -8.319 -0.6357 -0.3933 fixed
count_birth_order3/5+ -0.4626 0.0607 -7.622 -0.5816 -0.3437 fixed
count_birth_order4/5+ -0.351 0.05758 -6.095 -0.4638 -0.2381 fixed
count_birth_order5/5+ -0.3139 0.05473 -5.734 -0.4211 -0.2066 fixed
count_birth_order5+/5+ -0.3993 0.04395 -9.085 -0.4854 -0.3132 fixed
sd_(Intercept).mother_pidlink 0.5154 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.5662 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 13603 13677 -6790 13581 NA NA NA
12 13598 13678 -6787 13574 7.074 1 0.007821
16 13581 13688 -6774 13549 24.75 4 0.00005645
26 13596 13771 -6772 13544 4.93 10 0.8958

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.746 0.3083 -21.88 -7.351 -6.142 fixed
poly(age, 3, raw = TRUE)1 0.7193 0.035 20.55 0.6507 0.7879 fixed
poly(age, 3, raw = TRUE)2 -0.02245 0.00125 -17.96 -0.0249 -0.02 fixed
poly(age, 3, raw = TRUE)3 0.0002249 0.00001418 15.85 0.0001971 0.0002526 fixed
male -0.07773 0.01774 -4.381 -0.1125 -0.04296 fixed
sibling_count3 0.0339 0.03494 0.9704 -0.03457 0.1024 fixed
sibling_count4 -0.03127 0.03721 -0.8404 -0.1042 0.04166 fixed
sibling_count5 -0.08928 0.04042 -2.209 -0.1685 -0.01006 fixed
sibling_count5+ -0.2673 0.03522 -7.589 -0.3363 -0.1983 fixed
sd_(Intercept).mother_pidlink 0.5206 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.5663 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.75 0.3084 -21.89 -7.354 -6.145 fixed
birth_order 0.003115 0.005405 0.5763 -0.007478 0.01371 fixed
poly(age, 3, raw = TRUE)1 0.7191 0.035 20.55 0.6505 0.7877 fixed
poly(age, 3, raw = TRUE)2 -0.02245 0.00125 -17.96 -0.0249 -0.02 fixed
poly(age, 3, raw = TRUE)3 0.0002249 0.00001418 15.86 0.0001971 0.0002527 fixed
male -0.07789 0.01774 -4.39 -0.1127 -0.04311 fixed
sibling_count3 0.03222 0.03507 0.9189 -0.03651 0.101 fixed
sibling_count4 -0.03509 0.0378 -0.9283 -0.1092 0.039 fixed
sibling_count5 -0.09522 0.04172 -2.282 -0.177 -0.01344 fixed
sibling_count5+ -0.2794 0.04103 -6.809 -0.3598 -0.199 fixed
sd_(Intercept).mother_pidlink 0.521 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.5661 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.671 0.3092 -21.57 -7.277 -6.065 fixed
poly(age, 3, raw = TRUE)1 0.7127 0.03503 20.34 0.6441 0.7814 fixed
poly(age, 3, raw = TRUE)2 -0.02222 0.001251 -17.76 -0.02467 -0.01977 fixed
poly(age, 3, raw = TRUE)3 0.0002226 0.00001419 15.68 0.0001948 0.0002504 fixed
male -0.07837 0.01773 -4.421 -0.1131 -0.04362 fixed
sibling_count3 0.03597 0.03548 1.014 -0.03356 0.1055 fixed
sibling_count4 -0.03978 0.0387 -1.028 -0.1156 0.03606 fixed
sibling_count5 -0.1169 0.04304 -2.716 -0.2013 -0.03254 fixed
sibling_count5+ -0.2918 0.04193 -6.96 -0.374 -0.2096 fixed
birth_order_nonlinear2 -0.05895 0.02256 -2.613 -0.1032 -0.01474 fixed
birth_order_nonlinear3 -0.01423 0.0274 -0.5195 -0.06793 0.03946 fixed
birth_order_nonlinear4 0.03769 0.03338 1.129 -0.02773 0.1031 fixed
birth_order_nonlinear5 0.0732 0.04067 1.8 -0.006512 0.1529 fixed
birth_order_nonlinear5+ -0.01079 0.04033 -0.2675 -0.08984 0.06826 fixed
sd_(Intercept).mother_pidlink 0.5208 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.5654 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.652 0.3102 -21.44 -7.26 -6.044 fixed
poly(age, 3, raw = TRUE)1 0.7099 0.03512 20.22 0.6411 0.7788 fixed
poly(age, 3, raw = TRUE)2 -0.02212 0.001255 -17.63 -0.02458 -0.01966 fixed
poly(age, 3, raw = TRUE)3 0.0002213 0.00001423 15.55 0.0001934 0.0002492 fixed
male -0.07904 0.01775 -4.453 -0.1138 -0.04425 fixed
count_birth_order2/2 -0.04672 0.0446 -1.048 -0.1341 0.04069 fixed
count_birth_order1/3 0.05622 0.04275 1.315 -0.02757 0.14 fixed
count_birth_order2/3 -0.05503 0.04603 -1.196 -0.1452 0.03518 fixed
count_birth_order3/3 0.04289 0.05079 0.8444 -0.05666 0.1424 fixed
count_birth_order1/4 -0.07165 0.04984 -1.438 -0.1693 0.02603 fixed
count_birth_order2/4 -0.05084 0.05034 -1.01 -0.1495 0.04782 fixed
count_birth_order3/4 -0.06351 0.0545 -1.165 -0.1703 0.0433 fixed
count_birth_order4/4 0.009664 0.0564 0.1713 -0.1009 0.1202 fixed
count_birth_order1/5 -0.1223 0.05889 -2.077 -0.2377 -0.006873 fixed
count_birth_order2/5 -0.1555 0.06209 -2.504 -0.2772 -0.03379 fixed
count_birth_order3/5 -0.1429 0.06084 -2.349 -0.2621 -0.02364 fixed
count_birth_order4/5 -0.09629 0.06301 -1.528 -0.2198 0.02721 fixed
count_birth_order5/5 -0.007981 0.06277 -0.1271 -0.131 0.1151 fixed
count_birth_order1/5+ -0.2597 0.05562 -4.669 -0.3687 -0.1507 fixed
count_birth_order2/5+ -0.3699 0.05726 -6.461 -0.4822 -0.2577 fixed
count_birth_order3/5+ -0.2976 0.05583 -5.33 -0.407 -0.1881 fixed
count_birth_order4/5+ -0.2436 0.05381 -4.526 -0.3491 -0.1381 fixed
count_birth_order5/5+ -0.2379 0.05508 -4.318 -0.3458 -0.1299 fixed
count_birth_order5+/5+ -0.2999 0.04319 -6.943 -0.3845 -0.2152 fixed
sd_(Intercept).mother_pidlink 0.52 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.5661 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 13753 13827 -6866 13731 NA NA NA
12 13755 13836 -6865 13731 0.3298 1 0.5658
16 13747 13855 -6857 13715 16.17 4 0.002799
26 13760 13935 -6854 13708 6.573 10 0.765

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.758 0.313 -21.59 -7.371 -6.144 fixed
poly(age, 3, raw = TRUE)1 0.7212 0.03553 20.3 0.6516 0.7908 fixed
poly(age, 3, raw = TRUE)2 -0.02248 0.001269 -17.72 -0.02497 -0.02 fixed
poly(age, 3, raw = TRUE)3 0.0002253 0.0000144 15.65 0.0001971 0.0002536 fixed
male -0.07967 0.018 -4.426 -0.1149 -0.04439 fixed
sibling_count3 0.007354 0.03188 0.2306 -0.05514 0.06985 fixed
sibling_count4 -0.06524 0.03511 -1.858 -0.1341 0.00357 fixed
sibling_count5 -0.1584 0.04224 -3.75 -0.2412 -0.07564 fixed
sibling_count5+ -0.3751 0.03659 -10.25 -0.4469 -0.3034 fixed
sd_(Intercept).mother_pidlink 0.5174 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.568 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.769 0.3128 -21.64 -7.382 -6.156 fixed
birth_order 0.01578 0.006262 2.52 0.003506 0.02805 fixed
poly(age, 3, raw = TRUE)1 0.7196 0.03551 20.26 0.65 0.7892 fixed
poly(age, 3, raw = TRUE)2 -0.02244 0.001268 -17.69 -0.02493 -0.01995 fixed
poly(age, 3, raw = TRUE)3 0.0002255 0.00001439 15.67 0.0001973 0.0002538 fixed
male -0.08017 0.01799 -4.457 -0.1154 -0.04492 fixed
sibling_count3 -0.001263 0.03207 -0.03938 -0.06412 0.06159 fixed
sibling_count4 -0.08503 0.03598 -2.363 -0.1555 -0.01452 fixed
sibling_count5 -0.1895 0.04401 -4.305 -0.2757 -0.1032 fixed
sibling_count5+ -0.4382 0.04436 -9.88 -0.5252 -0.3513 fixed
sd_(Intercept).mother_pidlink 0.5182 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.5672 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.674 0.3137 -21.28 -7.289 -6.06 fixed
poly(age, 3, raw = TRUE)1 0.7136 0.03554 20.08 0.6439 0.7832 fixed
poly(age, 3, raw = TRUE)2 -0.02225 0.001269 -17.53 -0.02473 -0.01976 fixed
poly(age, 3, raw = TRUE)3 0.0002238 0.00001439 15.55 0.0001955 0.000252 fixed
male -0.08054 0.01796 -4.484 -0.1157 -0.04534 fixed
sibling_count3 0.00008121 0.03249 0.002499 -0.06361 0.06377 fixed
sibling_count4 -0.09508 0.03694 -2.574 -0.1675 -0.02268 fixed
sibling_count5 -0.2215 0.04531 -4.888 -0.3103 -0.1327 fixed
sibling_count5+ -0.4607 0.04527 -10.18 -0.5494 -0.372 fixed
birth_order_nonlinear2 -0.05633 0.02197 -2.563 -0.0994 -0.01326 fixed
birth_order_nonlinear3 0.01674 0.0274 0.611 -0.03696 0.07044 fixed
birth_order_nonlinear4 0.08775 0.03516 2.496 0.01884 0.1567 fixed
birth_order_nonlinear5 0.1372 0.04421 3.105 0.0506 0.2239 fixed
birth_order_nonlinear5+ 0.07675 0.04577 1.677 -0.01296 0.1665 fixed
sd_(Intercept).mother_pidlink 0.5176 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.5664 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -6.684 0.3148 -21.24 -7.301 -6.067 fixed
poly(age, 3, raw = TRUE)1 0.7138 0.03565 20.02 0.6439 0.7837 fixed
poly(age, 3, raw = TRUE)2 -0.02226 0.001273 -17.48 -0.02476 -0.01977 fixed
poly(age, 3, raw = TRUE)3 0.000224 0.00001445 15.5 0.0001957 0.0002523 fixed
male -0.07992 0.01799 -4.444 -0.1152 -0.04467 fixed
count_birth_order2/2 -0.02734 0.03915 -0.6983 -0.1041 0.0494 fixed
count_birth_order1/3 0.03303 0.03893 0.8485 -0.04327 0.1093 fixed
count_birth_order2/3 -0.06894 0.04228 -1.63 -0.1518 0.01393 fixed
count_birth_order3/3 0.008928 0.04599 0.1941 -0.08122 0.09907 fixed
count_birth_order1/4 -0.09519 0.04764 -1.998 -0.1886 -0.00181 fixed
count_birth_order2/4 -0.1545 0.04852 -3.184 -0.2496 -0.05937 fixed
count_birth_order3/4 -0.0412 0.05053 -0.8153 -0.1402 0.05784 fixed
count_birth_order4/4 0.003472 0.05308 0.06542 -0.1006 0.1075 fixed
count_birth_order1/5 -0.2173 0.06431 -3.379 -0.3434 -0.09126 fixed
count_birth_order2/5 -0.2431 0.06969 -3.489 -0.3797 -0.1065 fixed
count_birth_order3/5 -0.1786 0.06617 -2.699 -0.3082 -0.04887 fixed
count_birth_order4/5 -0.1625 0.06474 -2.509 -0.2893 -0.03557 fixed
count_birth_order5/5 -0.06279 0.06813 -0.9216 -0.1963 0.07075 fixed
count_birth_order1/5+ -0.4742 0.06382 -7.43 -0.5993 -0.3491 fixed
count_birth_order2/5+ -0.4845 0.06341 -7.64 -0.6087 -0.3602 fixed
count_birth_order3/5+ -0.4592 0.06144 -7.474 -0.5796 -0.3388 fixed
count_birth_order4/5+ -0.3303 0.06021 -5.486 -0.4483 -0.2123 fixed
count_birth_order5/5+ -0.3212 0.05599 -5.737 -0.431 -0.2115 fixed
count_birth_order5+/5+ -0.3735 0.04488 -8.322 -0.4615 -0.2855 fixed
sd_(Intercept).mother_pidlink 0.5172 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.5669 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 13356 13430 -6667 13334 NA NA NA
12 13352 13432 -6664 13328 6.342 1 0.01179
16 13338 13446 -6653 13306 21.48 4 0.0002543
26 13353 13528 -6651 13301 4.93 10 0.8958

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Elementary missed

birthorder <- birthorder %>% mutate(outcome = Elementary_missed)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.0359 0.07117 -0.5044 -0.1754 0.1036 fixed
poly(age, 3, raw = TRUE)1 0.005098 0.008201 0.6217 -0.01098 0.02117 fixed
poly(age, 3, raw = TRUE)2 -0.0001344 0.0002977 -0.4516 -0.0007179 0.000449 fixed
poly(age, 3, raw = TRUE)3 0.000001175 0.00000341 0.3446 -0.000005508 0.000007858 fixed
male 0.01183 0.004139 2.857 0.003713 0.01994 fixed
sibling_count3 -0.004516 0.007623 -0.5924 -0.01946 0.01042 fixed
sibling_count4 0.002375 0.007854 0.3024 -0.01302 0.01777 fixed
sibling_count5 0.0005693 0.008131 0.07002 -0.01537 0.0165 fixed
sibling_count5+ 0.01343 0.006512 2.062 0.0006617 0.02619 fixed
sd_(Intercept).mother_pidlink 0.03164 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1769 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.03567 0.07116 -0.5013 -0.1751 0.1038 fixed
birth_order 0.001618 0.000855 1.893 -0.00005746 0.003294 fixed
poly(age, 3, raw = TRUE)1 0.004764 0.008202 0.5809 -0.01131 0.02084 fixed
poly(age, 3, raw = TRUE)2 -0.0001223 0.0002977 -0.4107 -0.0007058 0.0004612 fixed
poly(age, 3, raw = TRUE)3 0.000001066 0.00000341 0.3126 -0.000005617 0.000007748 fixed
male 0.01187 0.004138 2.869 0.003762 0.01998 fixed
sibling_count3 -0.005119 0.007628 -0.6711 -0.02007 0.009832 fixed
sibling_count4 0.0008348 0.007895 0.1057 -0.01464 0.01631 fixed
sibling_count5 -0.001974 0.008239 -0.2396 -0.01812 0.01417 fixed
sibling_count5+ 0.005909 0.007626 0.7748 -0.009038 0.02086 fixed
sd_(Intercept).mother_pidlink 0.03155 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1769 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.03677 0.07132 -0.5155 -0.1766 0.103 fixed
poly(age, 3, raw = TRUE)1 0.005009 0.008207 0.6103 -0.01108 0.0211 fixed
poly(age, 3, raw = TRUE)2 -0.0001286 0.000298 -0.4316 -0.0007126 0.0004554 fixed
poly(age, 3, raw = TRUE)3 0.0000011 0.000003412 0.3224 -0.000005588 0.000007788 fixed
male 0.01198 0.00414 2.893 0.003862 0.02009 fixed
sibling_count3 -0.004969 0.007769 -0.6396 -0.0202 0.01026 fixed
sibling_count4 0.004068 0.008262 0.4924 -0.01213 0.02026 fixed
sibling_count5 0.001791 0.008784 0.2039 -0.01543 0.01901 fixed
sibling_count5+ 0.009914 0.008273 1.198 -0.006302 0.02613 fixed
birth_order_nonlinear2 0.002091 0.006096 0.3431 -0.009856 0.01404 fixed
birth_order_nonlinear3 0.002502 0.007308 0.3424 -0.01182 0.01683 fixed
birth_order_nonlinear4 -0.01034 0.008412 -1.229 -0.02682 0.006152 fixed
birth_order_nonlinear5 0.001859 0.009411 0.1975 -0.01659 0.0203 fixed
birth_order_nonlinear5+ 0.009152 0.007986 1.146 -0.006501 0.02481 fixed
sd_(Intercept).mother_pidlink 0.03137 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1769 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.04214 0.07141 -0.5902 -0.1821 0.09781 fixed
poly(age, 3, raw = TRUE)1 0.005408 0.008208 0.6588 -0.01068 0.02149 fixed
poly(age, 3, raw = TRUE)2 -0.0001434 0.000298 -0.4811 -0.0007274 0.0004407 fixed
poly(age, 3, raw = TRUE)3 0.000001267 0.000003413 0.3713 -0.000005422 0.000007956 fixed
male 0.01215 0.00414 2.935 0.004036 0.02027 fixed
count_birth_order2/2 0.007844 0.0116 0.6762 -0.01489 0.03058 fixed
count_birth_order1/3 0.001005 0.01024 0.09822 -0.01906 0.02107 fixed
count_birth_order2/3 -0.003272 0.01139 -0.2873 -0.02559 0.01905 fixed
count_birth_order3/3 -0.005575 0.01312 -0.425 -0.03129 0.02014 fixed
count_birth_order1/4 0.006385 0.01209 0.5281 -0.01731 0.03008 fixed
count_birth_order2/4 0.01971 0.01252 1.574 -0.004833 0.04425 fixed
count_birth_order3/4 -0.01251 0.01353 -0.9247 -0.03903 0.01401 fixed
count_birth_order4/4 0.002527 0.01402 0.1803 -0.02495 0.03 fixed
count_birth_order1/5 -0.008057 0.01393 -0.5784 -0.03536 0.01925 fixed
count_birth_order2/5 0.008299 0.01493 0.5557 -0.02097 0.03757 fixed
count_birth_order3/5 0.01165 0.01465 0.795 -0.01707 0.04037 fixed
count_birth_order4/5 0.004803 0.01549 0.3101 -0.02556 0.03517 fixed
count_birth_order5/5 0.001672 0.01482 0.1128 -0.02737 0.03071 fixed
count_birth_order1/5+ 0.01773 0.01285 1.38 -0.007455 0.04291 fixed
count_birth_order2/5+ -0.001829 0.01301 -0.1406 -0.02733 0.02367 fixed
count_birth_order3/5+ 0.03193 0.01249 2.557 0.007453 0.05641 fixed
count_birth_order4/5+ -0.007931 0.01181 -0.6717 -0.03107 0.01521 fixed
count_birth_order5/5+ 0.0157 0.01134 1.385 -0.006525 0.03793 fixed
count_birth_order5+/5+ 0.02106 0.008272 2.546 0.004846 0.03727 fixed
sd_(Intercept).mother_pidlink 0.0319 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1768 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -4483 -4407 2252 -4505 NA NA NA
12 -4485 -4401 2254 -4509 3.587 1 0.05823
16 -4478 -4367 2255 -4510 1.743 4 0.7829
26 -4472 -4292 2262 -4524 13.98 10 0.1738

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.1442 0.1069 -1.349 -0.3538 0.06538 fixed
poly(age, 3, raw = TRUE)1 0.01976 0.01294 1.527 -0.005603 0.04512 fixed
poly(age, 3, raw = TRUE)2 -0.0007469 0.0004952 -1.508 -0.001718 0.0002237 fixed
poly(age, 3, raw = TRUE)3 0.000009176 0.000005979 1.535 -0.000002543 0.0000209 fixed
male 0.006646 0.005295 1.255 -0.003731 0.01702 fixed
sibling_count3 -0.003274 0.007264 -0.4507 -0.01751 0.01096 fixed
sibling_count4 -0.004942 0.008107 -0.6095 -0.02083 0.01095 fixed
sibling_count5 0.002172 0.009593 0.2265 -0.01663 0.02097 fixed
sibling_count5+ 0.02326 0.00863 2.696 0.006348 0.04018 fixed
sd_(Intercept).mother_pidlink 0 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1667 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.145 0.107 -1.355 -0.3548 0.0647 fixed
birth_order 0.0004245 0.001896 0.2238 -0.003293 0.004142 fixed
poly(age, 3, raw = TRUE)1 0.0198 0.01294 1.53 -0.005568 0.04516 fixed
poly(age, 3, raw = TRUE)2 -0.0007491 0.0004954 -1.512 -0.00172 0.0002219 fixed
poly(age, 3, raw = TRUE)3 0.000009215 0.000005983 1.54 -0.00000251 0.00002094 fixed
male 0.006655 0.005296 1.257 -0.003724 0.01703 fixed
sibling_count3 -0.003473 0.007319 -0.4746 -0.01782 0.01087 fixed
sibling_count4 -0.005406 0.008369 -0.6459 -0.02181 0.011 fixed
sibling_count5 0.001394 0.0102 0.1366 -0.01861 0.02139 fixed
sibling_count5+ 0.02152 0.01163 1.85 -0.00128 0.04431 fixed
sd_(Intercept).mother_pidlink 0 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1667 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.1453 0.1071 -1.356 -0.3552 0.06464 fixed
poly(age, 3, raw = TRUE)1 0.0198 0.01295 1.529 -0.005579 0.04518 fixed
poly(age, 3, raw = TRUE)2 -0.0007478 0.0004957 -1.509 -0.001719 0.0002237 fixed
poly(age, 3, raw = TRUE)3 0.000009187 0.000005986 1.535 -0.000002546 0.00002092 fixed
male 0.006736 0.005298 1.271 -0.003648 0.01712 fixed
sibling_count3 -0.003312 0.007538 -0.4394 -0.01809 0.01146 fixed
sibling_count4 -0.000942 0.008859 -0.1063 -0.01831 0.01642 fixed
sibling_count5 0.007974 0.01095 0.7286 -0.01348 0.02943 fixed
sibling_count5+ 0.02341 0.01232 1.9 -0.0007407 0.04756 fixed
birth_order_nonlinear2 0.001382 0.006812 0.2029 -0.01197 0.01473 fixed
birth_order_nonlinear3 0.0001446 0.008687 0.01664 -0.01688 0.01717 fixed
birth_order_nonlinear4 -0.01699 0.01116 -1.523 -0.03886 0.004881 fixed
birth_order_nonlinear5 -0.007998 0.01468 -0.5448 -0.03677 0.02078 fixed
birth_order_nonlinear5+ 0.007723 0.01497 0.5159 -0.02162 0.03707 fixed
sd_(Intercept).mother_pidlink 0 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1667 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.143 0.1073 -1.332 -0.3533 0.06734 fixed
poly(age, 3, raw = TRUE)1 0.0197 0.01297 1.519 -0.005722 0.04513 fixed
poly(age, 3, raw = TRUE)2 -0.0007376 0.0004966 -1.485 -0.001711 0.0002357 fixed
poly(age, 3, raw = TRUE)3 0.000008982 0.000005999 1.497 -0.000002776 0.00002074 fixed
male 0.006549 0.005306 1.234 -0.00385 0.01695 fixed
count_birth_order2/2 -0.006425 0.01127 -0.5701 -0.02852 0.01566 fixed
count_birth_order1/3 -0.01274 0.009904 -1.286 -0.03215 0.006674 fixed
count_birth_order2/3 0.001926 0.01084 0.1776 -0.01933 0.02318 fixed
count_birth_order3/3 -0.001935 0.01213 -0.1595 -0.02572 0.02185 fixed
count_birth_order1/4 -0.000406 0.01319 -0.03077 -0.02627 0.02545 fixed
count_birth_order2/4 0.002909 0.01367 0.2127 -0.02389 0.02971 fixed
count_birth_order3/4 -0.01698 0.0144 -1.179 -0.04521 0.01125 fixed
count_birth_order4/4 -0.01695 0.01423 -1.19 -0.04485 0.01095 fixed
count_birth_order1/5 0.005682 0.01916 0.2965 -0.03187 0.04324 fixed
count_birth_order2/5 -0.001842 0.02077 -0.08868 -0.04255 0.03887 fixed
count_birth_order3/5 0.004445 0.01888 0.2354 -0.03257 0.04146 fixed
count_birth_order4/5 -0.0002872 0.01775 -0.01618 -0.03508 0.0345 fixed
count_birth_order5/5 -0.007808 0.01825 -0.4279 -0.04357 0.02795 fixed
count_birth_order1/5+ 0.04048 0.02442 1.658 -0.007385 0.08835 fixed
count_birth_order2/5+ 0.003233 0.02367 0.1366 -0.04315 0.04962 fixed
count_birth_order3/5+ 0.04167 0.02177 1.914 -0.0009953 0.08434 fixed
count_birth_order4/5+ -0.01883 0.0199 -0.9464 -0.05783 0.02017 fixed
count_birth_order5/5+ 0.01747 0.01754 0.9955 -0.01692 0.05185 fixed
count_birth_order5+/5+ 0.02843 0.01179 2.412 0.005327 0.05154 fixed
sd_(Intercept).mother_pidlink 0 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1668 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -2946 -2877 1484 -2968 NA NA NA
12 -2944 -2869 1484 -2968 0.05022 1 0.8227
16 -2940 -2840 1486 -2972 4.006 4 0.4052
26 -2929 -2765 1490 -2981 8.27 10 0.6024

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.1486 0.1065 -1.395 -0.3573 0.06015 fixed
poly(age, 3, raw = TRUE)1 0.02007 0.01288 1.558 -0.00518 0.04533 fixed
poly(age, 3, raw = TRUE)2 -0.0007564 0.0004933 -1.533 -0.001723 0.0002105 fixed
poly(age, 3, raw = TRUE)3 0.000009311 0.000005958 1.563 -0.000002366 0.00002099 fixed
male 0.006283 0.005263 1.194 -0.004032 0.0166 fixed
sibling_count3 0.00281 0.007829 0.3589 -0.01253 0.01815 fixed
sibling_count4 -0.005559 0.008443 -0.6585 -0.02211 0.01099 fixed
sibling_count5 -0.005294 0.009292 -0.5697 -0.02351 0.01292 fixed
sibling_count5+ 0.01816 0.008216 2.21 0.002056 0.03426 fixed
sd_(Intercept).mother_pidlink 0 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1666 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.1523 0.1065 -1.429 -0.3611 0.05653 fixed
birth_order 0.001729 0.001583 1.092 -0.001374 0.004832 fixed
poly(age, 3, raw = TRUE)1 0.02027 0.01289 1.573 -0.004982 0.04553 fixed
poly(age, 3, raw = TRUE)2 -0.0007663 0.0004934 -1.553 -0.001733 0.0002007 fixed
poly(age, 3, raw = TRUE)3 0.000009479 0.00000596 1.59 -0.000002202 0.00002116 fixed
male 0.006291 0.005263 1.195 -0.004024 0.01661 fixed
sibling_count3 0.00197 0.007866 0.2504 -0.01345 0.01739 fixed
sibling_count4 -0.007413 0.008611 -0.8609 -0.02429 0.009465 fixed
sibling_count5 -0.008217 0.00967 -0.8498 -0.02717 0.01074 fixed
sibling_count5+ 0.01149 0.01024 1.122 -0.008579 0.03155 fixed
sd_(Intercept).mother_pidlink 0 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1666 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.1515 0.1067 -1.42 -0.3606 0.05755 fixed
poly(age, 3, raw = TRUE)1 0.02036 0.0129 1.579 -0.004919 0.04564 fixed
poly(age, 3, raw = TRUE)2 -0.0007676 0.0004938 -1.554 -0.001735 0.0002004 fixed
poly(age, 3, raw = TRUE)3 0.000009457 0.000005966 1.585 -0.000002236 0.00002115 fixed
male 0.006342 0.005267 1.204 -0.003982 0.01667 fixed
sibling_count3 0.002582 0.008089 0.3193 -0.01327 0.01844 fixed
sibling_count4 -0.004501 0.009063 -0.4966 -0.02226 0.01326 fixed
sibling_count5 -0.003395 0.01036 -0.3278 -0.0237 0.01691 fixed
sibling_count5+ 0.01573 0.01075 1.464 -0.005334 0.0368 fixed
birth_order_nonlinear2 0.001636 0.006938 0.2358 -0.01196 0.01523 fixed
birth_order_nonlinear3 0.0009229 0.008591 0.1074 -0.01592 0.01776 fixed
birth_order_nonlinear4 -0.005306 0.01049 -0.506 -0.02586 0.01525 fixed
birth_order_nonlinear5 -0.003994 0.0136 -0.2938 -0.03064 0.02265 fixed
birth_order_nonlinear5+ 0.009254 0.0127 0.7289 -0.01563 0.03414 fixed
sd_(Intercept).mother_pidlink 0.00000008748 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1666 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.1472 0.1069 -1.377 -0.3566 0.06224 fixed
poly(age, 3, raw = TRUE)1 0.02004 0.01292 1.552 -0.005274 0.04535 fixed
poly(age, 3, raw = TRUE)2 -0.000755 0.0004946 -1.527 -0.001724 0.0002143 fixed
poly(age, 3, raw = TRUE)3 0.000009301 0.000005975 1.557 -0.000002408 0.00002101 fixed
male 0.006449 0.005273 1.223 -0.003887 0.01678 fixed
count_birth_order2/2 -0.003815 0.01228 -0.3108 -0.02788 0.02025 fixed
count_birth_order1/3 -0.003669 0.01069 -0.3433 -0.02461 0.01728 fixed
count_birth_order2/3 0.007273 0.0116 0.6268 -0.01547 0.03001 fixed
count_birth_order3/3 0.003067 0.01296 0.2366 -0.02234 0.02847 fixed
count_birth_order1/4 0.0007049 0.01346 0.05236 -0.02568 0.02709 fixed
count_birth_order2/4 0.001997 0.01358 0.147 -0.02462 0.02861 fixed
count_birth_order3/4 -0.02048 0.01513 -1.354 -0.05014 0.009174 fixed
count_birth_order4/4 -0.01489 0.01471 -1.012 -0.04372 0.01394 fixed
count_birth_order1/5 -0.01108 0.01672 -0.6626 -0.04386 0.0217 fixed
count_birth_order2/5 -0.005183 0.01899 -0.2729 -0.0424 0.03204 fixed
count_birth_order3/5 -0.008834 0.01738 -0.5082 -0.0429 0.02523 fixed
count_birth_order4/5 -0.008066 0.0178 -0.4531 -0.04296 0.02683 fixed
count_birth_order5/5 0.0008635 0.01743 0.04954 -0.0333 0.03502 fixed
count_birth_order1/5+ 0.0153 0.01781 0.8591 -0.0196 0.05021 fixed
count_birth_order2/5+ -0.006775 0.01984 -0.3415 -0.04566 0.03211 fixed
count_birth_order3/5+ 0.04067 0.01842 2.208 0.004567 0.07677 fixed
count_birth_order4/5+ 0.01091 0.01658 0.6578 -0.02159 0.0434 fixed
count_birth_order5/5+ -0.0003145 0.01759 -0.01788 -0.03479 0.03417 fixed
count_birth_order5+/5+ 0.02317 0.01124 2.062 0.001146 0.0452 fixed
sd_(Intercept).mother_pidlink 0.00000002179 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1667 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -2986 -2917 1504 -3008 NA NA NA
12 -2985 -2910 1505 -3009 1.196 1 0.2742
16 -2977 -2877 1505 -3009 0.2612 4 0.9922
26 -2965 -2802 1509 -3017 7.849 10 0.6436

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.1524 0.1069 -1.425 -0.362 0.05723 fixed
poly(age, 3, raw = TRUE)1 0.02052 0.01293 1.587 -0.004829 0.04587 fixed
poly(age, 3, raw = TRUE)2 -0.0007638 0.0004946 -1.544 -0.001733 0.0002056 fixed
poly(age, 3, raw = TRUE)3 0.0000093 0.000005968 1.558 -0.000002398 0.000021 fixed
male 0.004826 0.005319 0.9073 -0.005599 0.01525 fixed
sibling_count3 -0.001595 0.00716 -0.2227 -0.01563 0.01244 fixed
sibling_count4 -0.007724 0.008028 -0.9621 -0.02346 0.008011 fixed
sibling_count5 0.002436 0.009835 0.2477 -0.01684 0.02171 fixed
sibling_count5+ 0.02079 0.008775 2.37 0.003594 0.03799 fixed
sd_(Intercept).mother_pidlink 0 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1656 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.1561 0.107 -1.459 -0.3658 0.05362 fixed
birth_order 0.002012 0.001944 1.035 -0.001798 0.005823 fixed
poly(age, 3, raw = TRUE)1 0.02069 0.01293 1.599 -0.004663 0.04604 fixed
poly(age, 3, raw = TRUE)2 -0.0007734 0.0004947 -1.563 -0.001743 0.0001961 fixed
poly(age, 3, raw = TRUE)3 0.00000948 0.000005971 1.588 -0.000002222 0.00002118 fixed
male 0.004883 0.005319 0.9181 -0.005542 0.01531 fixed
sibling_count3 -0.002519 0.007215 -0.3492 -0.01666 0.01162 fixed
sibling_count4 -0.00991 0.008301 -1.194 -0.02618 0.006361 fixed
sibling_count5 -0.001106 0.01041 -0.1062 -0.02152 0.0193 fixed
sibling_count5+ 0.01253 0.01187 1.056 -0.01073 0.03578 fixed
sd_(Intercept).mother_pidlink 0.00000009768 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1656 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.1555 0.1071 -1.451 -0.3654 0.0545 fixed
poly(age, 3, raw = TRUE)1 0.02077 0.01294 1.605 -0.004599 0.04613 fixed
poly(age, 3, raw = TRUE)2 -0.0007756 0.000495 -1.567 -0.001746 0.0001947 fixed
poly(age, 3, raw = TRUE)3 0.000009493 0.000005975 1.589 -0.000002219 0.0000212 fixed
male 0.004933 0.005323 0.9269 -0.005499 0.01537 fixed
sibling_count3 -0.002476 0.007436 -0.333 -0.01705 0.0121 fixed
sibling_count4 -0.006352 0.008807 -0.7213 -0.02361 0.01091 fixed
sibling_count5 0.005045 0.01113 0.4534 -0.01677 0.02686 fixed
sibling_count5+ 0.0149 0.01266 1.177 -0.009919 0.03972 fixed
birth_order_nonlinear2 0.003457 0.006762 0.5112 -0.009797 0.01671 fixed
birth_order_nonlinear3 0.004017 0.008668 0.4635 -0.01297 0.02101 fixed
birth_order_nonlinear4 -0.009119 0.01146 -0.7957 -0.03158 0.01334 fixed
birth_order_nonlinear5 -0.003768 0.0153 -0.2462 -0.03376 0.02623 fixed
birth_order_nonlinear5+ 0.01701 0.01547 1.1 -0.01331 0.04734 fixed
sd_(Intercept).mother_pidlink 0 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1656 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.1605 0.1074 -1.495 -0.371 0.04999 fixed
poly(age, 3, raw = TRUE)1 0.02159 0.01297 1.664 -0.003843 0.04702 fixed
poly(age, 3, raw = TRUE)2 -0.0008016 0.0004963 -1.615 -0.001774 0.0001712 fixed
poly(age, 3, raw = TRUE)3 0.000009745 0.000005992 1.626 -0.000001999 0.00002149 fixed
male 0.005014 0.005333 0.9402 -0.005438 0.01547 fixed
count_birth_order2/2 -0.004404 0.0109 -0.404 -0.02577 0.01696 fixed
count_birth_order1/3 -0.01262 0.009776 -1.291 -0.03178 0.00654 fixed
count_birth_order2/3 0.003637 0.01089 0.334 -0.01771 0.02498 fixed
count_birth_order3/3 0.005645 0.01197 0.4714 -0.01782 0.02911 fixed
count_birth_order1/4 -0.004056 0.01325 -0.3062 -0.03002 0.02191 fixed
count_birth_order2/4 -0.001733 0.01367 -0.1268 -0.02852 0.02505 fixed
count_birth_order3/4 -0.01717 0.01406 -1.222 -0.04472 0.01038 fixed
count_birth_order4/4 -0.01619 0.01435 -1.128 -0.04431 0.01194 fixed
count_birth_order1/5 0.006246 0.01912 0.3266 -0.03123 0.04372 fixed
count_birth_order2/5 0.0009551 0.02176 0.04389 -0.0417 0.04361 fixed
count_birth_order3/5 0.009442 0.01987 0.4752 -0.0295 0.04839 fixed
count_birth_order4/5 0.003884 0.01857 0.2091 -0.03252 0.04029 fixed
count_birth_order5/5 -0.01661 0.0193 -0.8607 -0.05445 0.02122 fixed
count_birth_order1/5+ 0.02319 0.02464 0.9411 -0.02511 0.07149 fixed
count_birth_order2/5+ 0.005438 0.02432 0.2236 -0.04223 0.0531 fixed
count_birth_order3/5+ 0.01683 0.02315 0.7271 -0.02854 0.0622 fixed
count_birth_order4/5+ -0.01712 0.02153 -0.7949 -0.05932 0.02509 fixed
count_birth_order5/5+ 0.02131 0.01811 1.177 -0.01418 0.0568 fixed
count_birth_order5+/5+ 0.02908 0.01199 2.425 0.005578 0.05258 fixed
sd_(Intercept).mother_pidlink 0.000000008555 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1657 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -2935 -2866 1478 -2957 NA NA NA
12 -2934 -2859 1479 -2958 1.074 1 0.3001
16 -2928 -2828 1480 -2960 2.375 4 0.6672
26 -2915 -2752 1483 -2967 6.831 10 0.7413

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Elementary worked

birthorder <- birthorder %>% mutate(outcome = Elementary_worked)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.07964 0.08885 -0.8964 -0.2538 0.0945 fixed
poly(age, 3, raw = TRUE)1 0.00962 0.01022 0.9411 -0.01041 0.02966 fixed
poly(age, 3, raw = TRUE)2 -0.0003134 0.0003703 -0.8464 -0.001039 0.0004123 fixed
poly(age, 3, raw = TRUE)3 0.000004026 0.00000423 0.9519 -0.000004264 0.00001232 fixed
male 0.03192 0.005231 6.103 0.02167 0.04218 fixed
sibling_count3 -0.008452 0.009849 -0.8582 -0.02776 0.01085 fixed
sibling_count4 0.00381 0.01018 0.3744 -0.01613 0.02375 fixed
sibling_count5 0.01119 0.01059 1.057 -0.009562 0.03194 fixed
sibling_count5+ 0.03482 0.008449 4.122 0.01826 0.05138 fixed
sd_(Intercept).mother_pidlink 0.08269 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2132 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.07986 0.08885 -0.8988 -0.254 0.09429 fixed
birth_order 0.0009563 0.001108 0.8631 -0.001215 0.003128 fixed
poly(age, 3, raw = TRUE)1 0.009444 0.01022 0.9236 -0.0106 0.02948 fixed
poly(age, 3, raw = TRUE)2 -0.0003064 0.0003704 -0.8273 -0.001032 0.0004195 fixed
poly(age, 3, raw = TRUE)3 0.000003962 0.00000423 0.9364 -0.00000433 0.00001225 fixed
male 0.03195 0.005231 6.108 0.0217 0.04221 fixed
sibling_count3 -0.008806 0.009856 -0.8934 -0.02812 0.01051 fixed
sibling_count4 0.002898 0.01023 0.2833 -0.01715 0.02295 fixed
sibling_count5 0.009665 0.01073 0.9007 -0.01137 0.0307 fixed
sibling_count5+ 0.03038 0.009896 3.069 0.01098 0.04977 fixed
sd_(Intercept).mother_pidlink 0.08257 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2133 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.0841 0.08913 -0.9436 -0.2588 0.09058 fixed
poly(age, 3, raw = TRUE)1 0.009802 0.01024 0.9574 -0.01026 0.02987 fixed
poly(age, 3, raw = TRUE)2 -0.0003176 0.0003709 -0.8565 -0.001044 0.0004092 fixed
poly(age, 3, raw = TRUE)3 0.000004062 0.000004236 0.9589 -0.00000424 0.00001236 fixed
male 0.03209 0.005232 6.133 0.02184 0.04234 fixed
sibling_count3 -0.008324 0.01002 -0.8303 -0.02797 0.01132 fixed
sibling_count4 0.006027 0.01067 0.5647 -0.01489 0.02695 fixed
sibling_count5 0.01112 0.01139 0.9759 -0.01121 0.03344 fixed
sibling_count5+ 0.0318 0.01067 2.979 0.01088 0.05271 fixed
birth_order_nonlinear2 0.006231 0.007625 0.8172 -0.008714 0.02118 fixed
birth_order_nonlinear3 0.0001958 0.009161 0.02138 -0.01776 0.01815 fixed
birth_order_nonlinear4 -0.008989 0.01056 -0.8511 -0.02969 0.01171 fixed
birth_order_nonlinear5 0.01249 0.01183 1.056 -0.01069 0.03567 fixed
birth_order_nonlinear5+ 0.007857 0.01017 0.7724 -0.01208 0.02779 fixed
sd_(Intercept).mother_pidlink 0.08259 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2133 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.08246 0.08926 -0.9238 -0.2574 0.09248 fixed
poly(age, 3, raw = TRUE)1 0.009748 0.01024 0.952 -0.01032 0.02982 fixed
poly(age, 3, raw = TRUE)2 -0.0003177 0.000371 -0.8564 -0.001045 0.0004094 fixed
poly(age, 3, raw = TRUE)3 0.000004095 0.000004238 0.9664 -0.00000421 0.0000124 fixed
male 0.03227 0.005234 6.167 0.02202 0.04253 fixed
count_birth_order2/2 0.00326 0.01454 0.2242 -0.02524 0.03176 fixed
count_birth_order1/3 -0.004176 0.013 -0.3212 -0.02966 0.0213 fixed
count_birth_order2/3 0.0007817 0.01446 0.05408 -0.02755 0.02912 fixed
count_birth_order3/3 -0.02654 0.01664 -1.595 -0.05916 0.006077 fixed
count_birth_order1/4 0.001379 0.01533 0.08991 -0.02868 0.03143 fixed
count_birth_order2/4 0.01607 0.01589 1.011 -0.01507 0.04722 fixed
count_birth_order3/4 0.001507 0.01716 0.08783 -0.03213 0.03514 fixed
count_birth_order4/4 -0.00101 0.01778 -0.05678 -0.03586 0.03384 fixed
count_birth_order1/5 0.02121 0.01767 1.2 -0.01343 0.05585 fixed
count_birth_order2/5 0.008517 0.01889 0.451 -0.0285 0.04553 fixed
count_birth_order3/5 -0.004503 0.01857 -0.2426 -0.04089 0.03189 fixed
count_birth_order4/5 0.01404 0.01963 0.7155 -0.02443 0.05252 fixed
count_birth_order5/5 0.02075 0.01878 1.105 -0.01607 0.05756 fixed
count_birth_order1/5+ 0.01348 0.01626 0.829 -0.01839 0.04536 fixed
count_birth_order2/5+ 0.03364 0.01647 2.043 0.001366 0.06592 fixed
count_birth_order3/5+ 0.05796 0.0158 3.669 0.027 0.08893 fixed
count_birth_order4/5+ 0.01383 0.01495 0.9249 -0.01548 0.04314 fixed
count_birth_order5/5+ 0.04418 0.01434 3.08 0.01606 0.07229 fixed
count_birth_order5+/5+ 0.03869 0.01059 3.654 0.01793 0.05944 fixed
sd_(Intercept).mother_pidlink 0.08253 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2133 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -891.6 -815.4 456.8 -913.6 NA NA NA
12 -890.4 -807.2 457.2 -914.4 0.7467 1 0.3875
16 -885.9 -775 458.9 -917.9 3.481 4 0.4807
26 -877 -696.8 464.5 -929 11.14 10 0.3467

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.01121 0.12 0.09338 -0.224 0.2464 fixed
poly(age, 3, raw = TRUE)1 -0.0008552 0.01454 -0.05884 -0.02934 0.02763 fixed
poly(age, 3, raw = TRUE)2 0.0000577 0.0005569 0.1036 -0.001034 0.001149 fixed
poly(age, 3, raw = TRUE)3 -0.000001182 0.000006732 -0.1755 -0.00001438 0.00001201 fixed
male 0.02491 0.005955 4.183 0.01324 0.03658 fixed
sibling_count3 0.003041 0.008421 0.3611 -0.01346 0.01955 fixed
sibling_count4 0.02978 0.00946 3.148 0.01124 0.04832 fixed
sibling_count5 0.02227 0.01122 1.984 0.0002712 0.04427 fixed
sibling_count5+ 0.0533 0.01018 5.234 0.03334 0.07327 fixed
sd_(Intercept).mother_pidlink 0.07007 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.175 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.01518 0.1201 0.1263 -0.2203 0.2506 fixed
birth_order -0.001798 0.002186 -0.8224 -0.006082 0.002487 fixed
poly(age, 3, raw = TRUE)1 -0.001049 0.01454 -0.07217 -0.02954 0.02745 fixed
poly(age, 3, raw = TRUE)2 0.00006671 0.000557 0.1198 -0.001025 0.001159 fixed
poly(age, 3, raw = TRUE)3 -0.000001341 0.000006735 -0.1991 -0.00001454 0.00001186 fixed
male 0.02488 0.005956 4.178 0.01321 0.03656 fixed
sibling_count3 0.003896 0.008485 0.4591 -0.01273 0.02053 fixed
sibling_count4 0.03179 0.00977 3.254 0.01264 0.05094 fixed
sibling_count5 0.02563 0.01194 2.146 0.002219 0.04903 fixed
sibling_count5+ 0.0608 0.01367 4.448 0.03401 0.0876 fixed
sd_(Intercept).mother_pidlink 0.07 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1751 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.01003 0.1203 0.0834 -0.2258 0.2459 fixed
poly(age, 3, raw = TRUE)1 -0.00082 0.01455 -0.05634 -0.02935 0.02771 fixed
poly(age, 3, raw = TRUE)2 0.00005832 0.0005576 0.1046 -0.001035 0.001151 fixed
poly(age, 3, raw = TRUE)3 -0.000001221 0.000006742 -0.181 -0.00001443 0.00001199 fixed
male 0.02492 0.00596 4.182 0.01324 0.0366 fixed
sibling_count3 0.002344 0.008719 0.2688 -0.01475 0.01943 fixed
sibling_count4 0.03282 0.0103 3.186 0.01263 0.053 fixed
sibling_count5 0.02701 0.01274 2.121 0.002048 0.05197 fixed
sibling_count5+ 0.05584 0.01442 3.871 0.02757 0.08411 fixed
birth_order_nonlinear2 0.001738 0.007499 0.2318 -0.01296 0.01644 fixed
birth_order_nonlinear3 0.003074 0.009668 0.318 -0.01587 0.02202 fixed
birth_order_nonlinear4 -0.01505 0.01247 -1.206 -0.03949 0.009398 fixed
birth_order_nonlinear5 -0.007201 0.01644 -0.4379 -0.03943 0.02503 fixed
birth_order_nonlinear5+ 0.001383 0.0171 0.0809 -0.03214 0.0349 fixed
sd_(Intercept).mother_pidlink 0.07028 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.175 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.02475 0.1206 0.2052 -0.2116 0.2611 fixed
poly(age, 3, raw = TRUE)1 -0.002739 0.01459 -0.1878 -0.03133 0.02585 fixed
poly(age, 3, raw = TRUE)2 0.0001345 0.0005589 0.2407 -0.0009609 0.00123 fixed
poly(age, 3, raw = TRUE)3 -0.000002188 0.000006758 -0.3238 -0.00001543 0.00001106 fixed
male 0.02456 0.005969 4.114 0.01286 0.03625 fixed
count_birth_order2/2 0.004317 0.01246 0.3465 -0.0201 0.02874 fixed
count_birth_order1/3 0.003408 0.01118 0.3048 -0.01851 0.02532 fixed
count_birth_order2/3 0.006332 0.01223 0.5179 -0.01763 0.03029 fixed
count_birth_order3/3 0.004064 0.01367 0.2972 -0.02274 0.03086 fixed
count_birth_order1/4 0.03974 0.01488 2.671 0.01058 0.06891 fixed
count_birth_order2/4 0.02834 0.0154 1.84 -0.001853 0.05853 fixed
count_birth_order3/4 0.02931 0.01621 1.808 -0.002461 0.06109 fixed
count_birth_order4/4 0.02599 0.01604 1.62 -0.005454 0.05743 fixed
count_birth_order1/5 0.00689 0.0216 0.319 -0.03545 0.04923 fixed
count_birth_order2/5 0.03487 0.02336 1.492 -0.01092 0.08066 fixed
count_birth_order3/5 0.0276 0.02126 1.298 -0.01407 0.06927 fixed
count_birth_order4/5 0.01066 0.01998 0.5339 -0.02849 0.04982 fixed
count_birth_order5/5 0.04122 0.02055 2.006 0.0009388 0.08151 fixed
count_birth_order1/5+ 0.07669 0.02748 2.791 0.02284 0.1305 fixed
count_birth_order2/5+ 0.05725 0.02663 2.15 0.005063 0.1094 fixed
count_birth_order3/5+ 0.09302 0.02448 3.799 0.04503 0.141 fixed
count_birth_order4/5+ 0.02899 0.02237 1.296 -0.01484 0.07283 fixed
count_birth_order5/5+ 0.0316 0.01974 1.601 -0.007081 0.07028 fixed
count_birth_order5+/5+ 0.05784 0.01354 4.272 0.0313 0.08438 fixed
sd_(Intercept).mother_pidlink 0.07014 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1751 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -1991 -1922 1007 -2013 NA NA NA
12 -1990 -1915 1007 -2014 0.6787 1 0.41
16 -1984 -1883 1008 -2016 1.76 4 0.7799
26 -1971 -1808 1012 -2023 7.601 10 0.6677

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.0002828 0.1195 0.002366 -0.234 0.2345 fixed
poly(age, 3, raw = TRUE)1 0.0006424 0.01447 0.04439 -0.02772 0.02901 fixed
poly(age, 3, raw = TRUE)2 -0.000002017 0.0005547 -0.003637 -0.001089 0.001085 fixed
poly(age, 3, raw = TRUE)3 -0.0000003936 0.000006707 -0.05869 -0.00001354 0.00001275 fixed
male 0.02545 0.005918 4.3 0.01385 0.03704 fixed
sibling_count3 -0.002779 0.009064 -0.3066 -0.02054 0.01499 fixed
sibling_count4 0.02307 0.009816 2.35 0.003827 0.0423 fixed
sibling_count5 0.02523 0.01087 2.321 0.003927 0.04653 fixed
sibling_count5+ 0.03975 0.009611 4.136 0.02091 0.05859 fixed
sd_(Intercept).mother_pidlink 0.07022 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1748 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.001055 0.1196 0.008822 -0.2334 0.2355 fixed
birth_order -0.0003297 0.001834 -0.1797 -0.003925 0.003265 fixed
poly(age, 3, raw = TRUE)1 0.000602 0.01448 0.04159 -0.02777 0.02897 fixed
poly(age, 3, raw = TRUE)2 -0.0000002517 0.0005548 -0.0004536 -0.001088 0.001087 fixed
poly(age, 3, raw = TRUE)3 -0.000000423 0.00000671 -0.06305 -0.00001357 0.00001273 fixed
male 0.02545 0.005918 4.299 0.01385 0.03705 fixed
sibling_count3 -0.002616 0.009111 -0.2872 -0.02047 0.01524 fixed
sibling_count4 0.02343 0.01002 2.338 0.003786 0.04307 fixed
sibling_count5 0.02579 0.01132 2.279 0.003612 0.04798 fixed
sibling_count5+ 0.04103 0.01196 3.43 0.01759 0.06447 fixed
sd_(Intercept).mother_pidlink 0.07023 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1748 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.00532 0.1198 -0.04442 -0.2401 0.2294 fixed
poly(age, 3, raw = TRUE)1 0.00107 0.01449 0.07384 -0.02733 0.02947 fixed
poly(age, 3, raw = TRUE)2 -0.00001789 0.0005553 -0.03221 -0.001106 0.001071 fixed
poly(age, 3, raw = TRUE)3 -0.0000001971 0.000006716 -0.02934 -0.00001336 0.00001297 fixed
male 0.02563 0.005921 4.329 0.01403 0.03724 fixed
sibling_count3 -0.002255 0.009346 -0.2413 -0.02057 0.01606 fixed
sibling_count4 0.0258 0.0105 2.456 0.005212 0.04639 fixed
sibling_count5 0.02825 0.01205 2.344 0.00463 0.05187 fixed
sibling_count5+ 0.03817 0.01249 3.057 0.0137 0.06264 fixed
birth_order_nonlinear2 0.00555 0.007641 0.7263 -0.009427 0.02053 fixed
birth_order_nonlinear3 -0.002217 0.009564 -0.2318 -0.02096 0.01653 fixed
birth_order_nonlinear4 -0.008889 0.01172 -0.7585 -0.03186 0.01408 fixed
birth_order_nonlinear5 0.00004574 0.01522 0.003004 -0.02979 0.02989 fixed
birth_order_nonlinear5+ 0.01021 0.01449 0.7042 -0.0182 0.03861 fixed
sd_(Intercept).mother_pidlink 0.07045 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1747 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.001061 0.12 0.008841 -0.2342 0.2363 fixed
poly(age, 3, raw = TRUE)1 0.0001726 0.01452 0.01189 -0.02828 0.02862 fixed
poly(age, 3, raw = TRUE)2 0.00001582 0.0005564 0.02843 -0.001075 0.001106 fixed
poly(age, 3, raw = TRUE)3 -0.0000005898 0.000006729 -0.08766 -0.00001378 0.0000126 fixed
male 0.02548 0.005931 4.297 0.01386 0.03711 fixed
count_birth_order2/2 0.008978 0.01357 0.6614 -0.01763 0.03558 fixed
count_birth_order1/3 0.003922 0.01206 0.3251 -0.01972 0.02757 fixed
count_birth_order2/3 0.0002864 0.01308 0.02189 -0.02535 0.02592 fixed
count_birth_order3/3 -0.007098 0.01461 -0.4859 -0.03573 0.02153 fixed
count_birth_order1/4 0.02279 0.01518 1.501 -0.006973 0.05255 fixed
count_birth_order2/4 0.02978 0.0153 1.946 -0.0002105 0.05978 fixed
count_birth_order3/4 0.01887 0.01703 1.108 -0.01451 0.05225 fixed
count_birth_order4/4 0.03166 0.01658 1.91 -0.0008367 0.06416 fixed
count_birth_order1/5 0.03335 0.01886 1.768 -0.003614 0.07031 fixed
count_birth_order2/5 0.04802 0.02126 2.259 0.006355 0.08968 fixed
count_birth_order3/5 0.02118 0.01957 1.083 -0.01716 0.05953 fixed
count_birth_order4/5 0.01393 0.02002 0.6957 -0.02531 0.05317 fixed
count_birth_order5/5 0.02655 0.01963 1.353 -0.01191 0.06502 fixed
count_birth_order1/5+ 0.02823 0.02007 1.407 -0.01111 0.06756 fixed
count_birth_order2/5+ 0.0452 0.02232 2.025 0.001454 0.08894 fixed
count_birth_order3/5+ 0.06224 0.02072 3.003 0.02162 0.1029 fixed
count_birth_order4/5+ 0.01788 0.01865 0.9585 -0.01868 0.05443 fixed
count_birth_order5/5+ 0.04218 0.01978 2.132 0.003409 0.08096 fixed
count_birth_order5+/5+ 0.04944 0.01289 3.835 0.02417 0.07471 fixed
sd_(Intercept).mother_pidlink 0.0702 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1749 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -2022 -1953 1022 -2044 NA NA NA
12 -2020 -1945 1022 -2044 0.03257 1 0.8568
16 -2015 -1914 1023 -2047 2.509 4 0.6429
26 -2000 -1837 1026 -2052 5.464 10 0.8581

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.01789 0.1209 0.148 -0.219 0.2548 fixed
poly(age, 3, raw = TRUE)1 -0.001561 0.01463 -0.1067 -0.03023 0.02711 fixed
poly(age, 3, raw = TRUE)2 0.00008934 0.0005601 0.1595 -0.001008 0.001187 fixed
poly(age, 3, raw = TRUE)3 -0.000001613 0.000006766 -0.2384 -0.00001487 0.00001165 fixed
male 0.02612 0.006023 4.337 0.01432 0.03793 fixed
sibling_count3 0.0007046 0.008364 0.08425 -0.01569 0.0171 fixed
sibling_count4 0.021 0.009442 2.224 0.002494 0.03951 fixed
sibling_count5 0.02668 0.01162 2.295 0.003894 0.04946 fixed
sibling_count5+ 0.05476 0.01044 5.245 0.03429 0.07522 fixed
sd_(Intercept).mother_pidlink 0.07021 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1751 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.02227 0.121 0.1841 -0.2148 0.2594 fixed
birth_order -0.002079 0.00225 -0.9241 -0.006489 0.002331 fixed
poly(age, 3, raw = TRUE)1 -0.001768 0.01463 -0.1208 -0.03044 0.02691 fixed
poly(age, 3, raw = TRUE)2 0.00009948 0.0005602 0.1776 -0.0009985 0.001197 fixed
poly(age, 3, raw = TRUE)3 -0.000001796 0.000006769 -0.2653 -0.00001506 0.00001147 fixed
male 0.02607 0.006024 4.328 0.01427 0.03788 fixed
sibling_count3 0.001673 0.008429 0.1985 -0.01485 0.01819 fixed
sibling_count4 0.02331 0.009766 2.387 0.004166 0.04245 fixed
sibling_count5 0.03041 0.0123 2.471 0.006292 0.05452 fixed
sibling_count5+ 0.06342 0.01403 4.52 0.03592 0.09091 fixed
sd_(Intercept).mother_pidlink 0.07014 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1751 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.0182 0.1212 0.1502 -0.2193 0.2557 fixed
poly(age, 3, raw = TRUE)1 -0.001587 0.01465 -0.1083 -0.0303 0.02712 fixed
poly(age, 3, raw = TRUE)2 0.00009018 0.0005608 0.1608 -0.001009 0.001189 fixed
poly(age, 3, raw = TRUE)3 -0.000001632 0.000006776 -0.2409 -0.00001491 0.00001165 fixed
male 0.02602 0.006028 4.317 0.01421 0.03784 fixed
sibling_count3 0.000491 0.008668 0.05664 -0.0165 0.01748 fixed
sibling_count4 0.02332 0.01032 2.261 0.003102 0.04354 fixed
sibling_count5 0.03303 0.01308 2.526 0.007402 0.05867 fixed
sibling_count5+ 0.0583 0.01491 3.911 0.02909 0.08751 fixed
birth_order_nonlinear2 0.0002169 0.007492 0.02895 -0.01447 0.0149 fixed
birth_order_nonlinear3 0.001036 0.009708 0.1067 -0.01799 0.02006 fixed
birth_order_nonlinear4 -0.01089 0.01288 -0.8453 -0.03615 0.01436 fixed
birth_order_nonlinear5 -0.019 0.01726 -1.101 -0.05282 0.01482 fixed
birth_order_nonlinear5+ 0.001732 0.01775 0.09756 -0.03306 0.03653 fixed
sd_(Intercept).mother_pidlink 0.07048 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.175 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.03478 0.1215 0.2863 -0.2034 0.2729 fixed
poly(age, 3, raw = TRUE)1 -0.003973 0.01469 -0.2706 -0.03276 0.02481 fixed
poly(age, 3, raw = TRUE)2 0.0001861 0.0005623 0.3309 -0.000916 0.001288 fixed
poly(age, 3, raw = TRUE)3 -0.000002854 0.000006796 -0.4199 -0.00001617 0.00001047 fixed
male 0.02537 0.006038 4.202 0.01354 0.0372 fixed
count_birth_order2/2 0.007706 0.01212 0.6361 -0.01604 0.03145 fixed
count_birth_order1/3 0.004014 0.01111 0.3613 -0.01776 0.02579 fixed
count_birth_order2/3 0.003202 0.01236 0.2591 -0.02102 0.02742 fixed
count_birth_order3/3 0.002463 0.01358 0.1813 -0.02416 0.02909 fixed
count_birth_order1/4 0.03037 0.01504 2.02 0.0009027 0.05984 fixed
count_birth_order2/4 0.02255 0.0155 1.455 -0.007823 0.05292 fixed
count_birth_order3/4 0.0205 0.01592 1.287 -0.01071 0.05171 fixed
count_birth_order4/4 0.02008 0.01627 1.234 -0.01181 0.05197 fixed
count_birth_order1/5 0.01887 0.0217 0.8697 -0.02365 0.06139 fixed
count_birth_order2/5 0.03535 0.02463 1.435 -0.01292 0.08362 fixed
count_birth_order3/5 0.03267 0.02251 1.451 -0.01145 0.07679 fixed
count_birth_order4/5 0.02488 0.02104 1.183 -0.01634 0.06611 fixed
count_birth_order5/5 0.03749 0.02189 1.713 -0.005409 0.08038 fixed
count_birth_order1/5+ 0.09729 0.02792 3.485 0.04258 0.152 fixed
count_birth_order2/5+ 0.04167 0.02754 1.513 -0.01231 0.09565 fixed
count_birth_order3/5+ 0.09556 0.0262 3.647 0.04421 0.1469 fixed
count_birth_order4/5+ 0.03713 0.02435 1.525 -0.01058 0.08485 fixed
count_birth_order5/5+ 0.02453 0.0205 1.197 -0.01565 0.06472 fixed
count_birth_order5+/5+ 0.06231 0.01387 4.493 0.03513 0.08949 fixed
sd_(Intercept).mother_pidlink 0.07055 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.175 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -1946 -1877 984.1 -1968 NA NA NA
12 -1945 -1870 984.5 -1969 0.857 1 0.3546
16 -1938 -1838 985.2 -1970 1.428 4 0.8393
26 -1927 -1764 989.3 -1979 8.208 10 0.6085

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Attended School

birthorder <- birthorder %>% mutate(outcome = attended_school)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.9681 0.01706 56.76 0.9347 1.002 fixed
poly(age, 3, raw = TRUE)1 0.002032 0.001622 1.253 -0.001146 0.00521 fixed
poly(age, 3, raw = TRUE)2 -0.00002256 0.00004725 -0.4774 -0.0001152 0.00007005 fixed
poly(age, 3, raw = TRUE)3 -0.0000005908 0.0000004301 -1.374 -0.000001434 0.0000002522 fixed
male 0.0045 0.00183 2.459 0.0009138 0.008086 fixed
sibling_count3 -0.0003612 0.00372 -0.09709 -0.007653 0.00693 fixed
sibling_count4 -0.0006508 0.003832 -0.1698 -0.008161 0.00686 fixed
sibling_count5 0.003564 0.004005 0.8899 -0.004285 0.01141 fixed
sibling_count5+ -0.003767 0.003134 -1.202 -0.00991 0.002375 fixed
sd_(Intercept).mother_pidlink 0.04 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1049 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.9682 0.01706 56.77 0.9348 1.002 fixed
birth_order 0.0003991 0.0003879 1.029 -0.0003612 0.001159 fixed
poly(age, 3, raw = TRUE)1 0.001915 0.001626 1.178 -0.001271 0.005101 fixed
poly(age, 3, raw = TRUE)2 -0.00001828 0.00004743 -0.3854 -0.0001113 0.00007469 fixed
poly(age, 3, raw = TRUE)3 -0.0000006301 0.0000004318 -1.459 -0.000001476 0.0000002162 fixed
male 0.00449 0.00183 2.454 0.0009035 0.008076 fixed
sibling_count3 -0.0004555 0.003722 -0.1224 -0.00775 0.006839 fixed
sibling_count4 -0.0009254 0.003842 -0.2409 -0.008455 0.006604 fixed
sibling_count5 0.003088 0.004032 0.7659 -0.004814 0.01099 fixed
sibling_count5+ -0.005271 0.003458 -1.524 -0.01205 0.001507 fixed
sd_(Intercept).mother_pidlink 0.04003 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1049 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.9654 0.0171 56.47 0.9319 0.9989 fixed
poly(age, 3, raw = TRUE)1 0.001958 0.001626 1.204 -0.001229 0.005144 fixed
poly(age, 3, raw = TRUE)2 -0.00002171 0.00004743 -0.4577 -0.0001147 0.00007126 fixed
poly(age, 3, raw = TRUE)3 -0.000000577 0.0000004318 -1.336 -0.000001423 0.0000002694 fixed
male 0.004508 0.001829 2.464 0.0009228 0.008093 fixed
sibling_count3 -0.0008396 0.003773 -0.2226 -0.008234 0.006554 fixed
sibling_count4 -0.001915 0.003947 -0.4851 -0.009651 0.005821 fixed
sibling_count5 0.002337 0.004181 0.5589 -0.005858 0.01053 fixed
sibling_count5+ -0.005684 0.003629 -1.566 -0.0128 0.00143 fixed
birth_order_nonlinear2 0.009478 0.002651 3.576 0.004282 0.01467 fixed
birth_order_nonlinear3 0.004998 0.00312 1.602 -0.001118 0.01111 fixed
birth_order_nonlinear4 0.00704 0.003569 1.972 0.00004392 0.01404 fixed
birth_order_nonlinear5 0.00423 0.00406 1.042 -0.003728 0.01219 fixed
birth_order_nonlinear5+ 0.006468 0.003369 1.92 -0.0001353 0.01307 fixed
sd_(Intercept).mother_pidlink 0.04003 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1049 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.9637 0.01717 56.11 0.93 0.9973 fixed
poly(age, 3, raw = TRUE)1 0.001951 0.001626 1.2 -0.001236 0.005139 fixed
poly(age, 3, raw = TRUE)2 -0.00002194 0.00004745 -0.4624 -0.0001149 0.00007106 fixed
poly(age, 3, raw = TRUE)3 -0.0000005714 0.0000004321 -1.322 -0.000001418 0.0000002755 fixed
male 0.004513 0.00183 2.466 0.0009264 0.0081 fixed
count_birth_order2/2 0.01457 0.005155 2.826 0.004466 0.02467 fixed
count_birth_order1/3 0.003496 0.004912 0.7117 -0.006132 0.01312 fixed
count_birth_order2/3 0.006761 0.005492 1.231 -0.004004 0.01753 fixed
count_birth_order3/3 0.006183 0.006121 1.01 -0.005814 0.01818 fixed
count_birth_order1/4 -0.001403 0.00556 -0.2522 -0.0123 0.009496 fixed
count_birth_order2/4 0.00841 0.005915 1.422 -0.003185 0.02 fixed
count_birth_order3/4 0.007888 0.006414 1.23 -0.004684 0.02046 fixed
count_birth_order4/4 0.00767 0.006728 1.14 -0.005515 0.02086 fixed
count_birth_order1/5 0.003475 0.006362 0.5463 -0.008993 0.01594 fixed
count_birth_order2/5 0.0146 0.006711 2.175 0.001442 0.02775 fixed
count_birth_order3/5 0.009876 0.006864 1.439 -0.003577 0.02333 fixed
count_birth_order4/5 0.01004 0.007349 1.367 -0.00436 0.02445 fixed
count_birth_order5/5 0.00893 0.007486 1.193 -0.005743 0.0236 fixed
count_birth_order1/5+ -0.002407 0.005119 -0.4702 -0.01244 0.007626 fixed
count_birth_order2/5+ 0.006208 0.005288 1.174 -0.004157 0.01657 fixed
count_birth_order3/5+ -0.0006546 0.005174 -0.1265 -0.0108 0.009487 fixed
count_birth_order4/5+ 0.003439 0.005098 0.6747 -0.006552 0.01343 fixed
count_birth_order5/5+ 0.0002875 0.005118 0.05618 -0.009743 0.01032 fixed
count_birth_order5+/5+ 0.002677 0.00402 0.6661 -0.005201 0.01056 fixed
sd_(Intercept).mother_pidlink 0.04003 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1049 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -22818 -22735 11420 -22840 NA NA NA
12 -22817 -22726 11421 -22841 1.058 1 0.3037
16 -22822 -22701 11427 -22854 12.65 4 0.01311
26 -22805 -22608 11429 -22857 3.176 10 0.977

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.011 0.02852 35.46 0.9553 1.067 fixed
poly(age, 3, raw = TRUE)1 -0.001947 0.003235 -0.6018 -0.008288 0.004394 fixed
poly(age, 3, raw = TRUE)2 0.0000824 0.0001155 0.7134 -0.000144 0.0003088 fixed
poly(age, 3, raw = TRUE)3 -0.000001031 0.000001309 -0.7881 -0.000003597 0.000001534 fixed
male 0.0001955 0.001662 0.1176 -0.003062 0.003453 fixed
sibling_count3 -0.0003751 0.002516 -0.1491 -0.005307 0.004557 fixed
sibling_count4 -0.001316 0.002683 -0.4904 -0.006575 0.003943 fixed
sibling_count5 -0.003354 0.003057 -1.097 -0.009345 0.002636 fixed
sibling_count5+ -0.006361 0.002651 -2.399 -0.01156 -0.001165 fixed
sd_(Intercept).mother_pidlink 0 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.0656 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.012 0.02852 35.48 0.956 1.068 fixed
birth_order -0.0007748 0.00054 -1.435 -0.001833 0.0002837 fixed
poly(age, 3, raw = TRUE)1 -0.001931 0.003235 -0.5969 -0.008272 0.00441 fixed
poly(age, 3, raw = TRUE)2 0.00008326 0.0001155 0.7209 -0.0001431 0.0003096 fixed
poly(age, 3, raw = TRUE)3 -0.00000107 0.000001309 -0.8174 -0.000003636 0.000001496 fixed
male 0.0002228 0.001662 0.1341 -0.003034 0.00348 fixed
sibling_count3 0.00000391 0.00253 0.001546 -0.004955 0.004962 fixed
sibling_count4 -0.0004532 0.002749 -0.1648 -0.005842 0.004936 fixed
sibling_count5 -0.001941 0.003211 -0.6046 -0.008235 0.004352 fixed
sibling_count5+ -0.003485 0.003324 -1.048 -0.009999 0.00303 fixed
sd_(Intercept).mother_pidlink 0 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.0656 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.015 0.02855 35.54 0.9589 1.071 fixed
poly(age, 3, raw = TRUE)1 -0.002183 0.003236 -0.6745 -0.008526 0.00416 fixed
poly(age, 3, raw = TRUE)2 0.00009152 0.0001155 0.7921 -0.0001349 0.000318 fixed
poly(age, 3, raw = TRUE)3 -0.000001145 0.00000131 -0.8742 -0.000003712 0.000001422 fixed
male 0.0001161 0.001662 0.06984 -0.003141 0.003373 fixed
sibling_count3 -0.0003986 0.00259 -0.1539 -0.005476 0.004678 fixed
sibling_count4 -0.00217 0.002874 -0.7549 -0.007803 0.003464 fixed
sibling_count5 -0.00477 0.0034 -1.403 -0.01143 0.001894 fixed
sibling_count5+ -0.006043 0.003442 -1.756 -0.01279 0.0007036 fixed
birth_order_nonlinear2 -0.005037 0.002194 -2.296 -0.009336 -0.0007372 fixed
birth_order_nonlinear3 -0.0001685 0.002691 -0.06262 -0.005444 0.005107 fixed
birth_order_nonlinear4 0.002693 0.003327 0.8095 -0.003828 0.009215 fixed
birth_order_nonlinear5 0.0002744 0.004181 0.06563 -0.00792 0.008469 fixed
birth_order_nonlinear5+ -0.004911 0.00408 -1.204 -0.01291 0.003085 fixed
sd_(Intercept).mother_pidlink 0 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.06558 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.014 0.02858 35.49 0.9585 1.071 fixed
poly(age, 3, raw = TRUE)1 -0.00219 0.00324 -0.6758 -0.00854 0.00416 fixed
poly(age, 3, raw = TRUE)2 0.00009155 0.0001157 0.7913 -0.0001352 0.0003183 fixed
poly(age, 3, raw = TRUE)3 -0.000001146 0.000001312 -0.8737 -0.000003717 0.000001425 fixed
male 0.0001459 0.001662 0.08779 -0.003111 0.003403 fixed
count_birth_order2/2 -0.003615 0.003935 -0.9188 -0.01133 0.004096 fixed
count_birth_order1/3 -0.0004489 0.003422 -0.1312 -0.007157 0.006259 fixed
count_birth_order2/3 -0.004972 0.003731 -1.333 -0.01228 0.00234 fixed
count_birth_order3/3 0.0009987 0.004173 0.2393 -0.00718 0.009177 fixed
count_birth_order1/4 -0.001972 0.004172 -0.4727 -0.01015 0.006204 fixed
count_birth_order2/4 -0.005412 0.004322 -1.252 -0.01388 0.003059 fixed
count_birth_order3/4 0.000761 0.00458 0.1662 -0.008216 0.009738 fixed
count_birth_order4/4 -0.003091 0.004762 -0.6492 -0.01242 0.006241 fixed
count_birth_order1/5 -0.005454 0.005692 -0.9582 -0.01661 0.005702 fixed
count_birth_order2/5 -0.02171 0.006134 -3.539 -0.03373 -0.009688 fixed
count_birth_order3/5 0.0004959 0.00569 0.08716 -0.01066 0.01165 fixed
count_birth_order4/5 0.0006201 0.005594 0.1109 -0.01034 0.01158 fixed
count_birth_order5/5 0.0007406 0.005808 0.1275 -0.01064 0.01212 fixed
count_birth_order1/5+ 0.0007324 0.00565 0.1296 -0.01034 0.01181 fixed
count_birth_order2/5+ -0.005395 0.005634 -0.9577 -0.01644 0.005647 fixed
count_birth_order3/5+ -0.01633 0.005546 -2.945 -0.0272 -0.005465 fixed
count_birth_order4/5+ 0.0004021 0.005225 0.07696 -0.009838 0.01064 fixed
count_birth_order5/5+ -0.00845 0.005 -1.69 -0.01825 0.00135 fixed
count_birth_order5+/5+ -0.01044 0.003679 -2.838 -0.01765 -0.003229 fixed
sd_(Intercept).mother_pidlink 0 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.06554 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -16291 -16217 8157 -16313 NA NA NA
12 -16291 -16210 8158 -16315 2.061 1 0.1511
16 -16291 -16183 8161 -16323 7.77 4 0.1004
26 -16288 -16113 8170 -16340 16.98 10 0.07483

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.012 0.02829 35.76 0.9561 1.067 fixed
poly(age, 3, raw = TRUE)1 -0.001995 0.003211 -0.6213 -0.008289 0.004299 fixed
poly(age, 3, raw = TRUE)2 0.00008284 0.0001147 0.7224 -0.0001419 0.0003076 fixed
poly(age, 3, raw = TRUE)3 -0.000001035 0.0000013 -0.7966 -0.000003583 0.000001512 fixed
male 0.0001628 0.001648 0.09879 -0.003067 0.003392 fixed
sibling_count3 -0.0002068 0.002719 -0.07605 -0.005537 0.005123 fixed
sibling_count4 -0.00003864 0.002839 -0.01361 -0.005603 0.005526 fixed
sibling_count5 -0.001917 0.003027 -0.6333 -0.007849 0.004016 fixed
sibling_count5+ -0.00446 0.002648 -1.684 -0.009651 0.0007303 fixed
sd_(Intercept).mother_pidlink 0 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.06532 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.012 0.02829 35.78 0.9569 1.068 fixed
birth_order -0.0006005 0.0004664 -1.288 -0.001515 0.0003136 fixed
poly(age, 3, raw = TRUE)1 -0.002002 0.003211 -0.6233 -0.008296 0.004292 fixed
poly(age, 3, raw = TRUE)2 0.00008416 0.0001147 0.7339 -0.0001406 0.0003089 fixed
poly(age, 3, raw = TRUE)3 -0.000001072 0.0000013 -0.8247 -0.00000362 0.000001476 fixed
male 0.000178 0.001648 0.108 -0.003051 0.003408 fixed
sibling_count3 0.00008331 0.002729 0.03053 -0.005265 0.005431 fixed
sibling_count4 0.0006008 0.002882 0.2085 -0.005048 0.00625 fixed
sibling_count5 -0.0009089 0.003126 -0.2907 -0.007036 0.005218 fixed
sibling_count5+ -0.002282 0.003142 -0.7263 -0.008441 0.003877 fixed
sd_(Intercept).mother_pidlink 0.00000000137 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.06532 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.015 0.02832 35.85 0.9599 1.071 fixed
poly(age, 3, raw = TRUE)1 -0.002241 0.003211 -0.6979 -0.008535 0.004053 fixed
poly(age, 3, raw = TRUE)2 0.00009272 0.0001147 0.8084 -0.0001321 0.0003175 fixed
poly(age, 3, raw = TRUE)3 -0.000001162 0.0000013 -0.8934 -0.00000371 0.000001387 fixed
male 0.00009342 0.001647 0.05671 -0.003135 0.003322 fixed
sibling_count3 -0.0001752 0.002785 -0.06292 -0.005633 0.005283 fixed
sibling_count4 -0.0005815 0.002996 -0.1941 -0.006453 0.00529 fixed
sibling_count5 -0.003249 0.003294 -0.9863 -0.009706 0.003208 fixed
sibling_count5+ -0.003948 0.003257 -1.212 -0.01033 0.002435 fixed
birth_order_nonlinear2 -0.005297 0.00222 -2.386 -0.009649 -0.000946 fixed
birth_order_nonlinear3 -0.0005044 0.002664 -0.1893 -0.005726 0.004717 fixed
birth_order_nonlinear4 0.001782 0.003203 0.5562 -0.004497 0.00806 fixed
birth_order_nonlinear5 0.001802 0.003933 0.4582 -0.005906 0.00951 fixed
birth_order_nonlinear5+ -0.005815 0.003619 -1.607 -0.01291 0.001278 fixed
sd_(Intercept).mother_pidlink 0 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.06529 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.015 0.02836 35.79 0.9593 1.071 fixed
poly(age, 3, raw = TRUE)1 -0.002231 0.003215 -0.694 -0.008532 0.00407 fixed
poly(age, 3, raw = TRUE)2 0.00009229 0.0001148 0.8037 -0.0001328 0.0003174 fixed
poly(age, 3, raw = TRUE)3 -0.000001156 0.000001302 -0.8876 -0.000003708 0.000001396 fixed
male 0.0001064 0.001648 0.06454 -0.003123 0.003336 fixed
count_birth_order2/2 -0.004276 0.004292 -0.9963 -0.01269 0.004136 fixed
count_birth_order1/3 -0.0003683 0.003692 -0.09976 -0.007605 0.006868 fixed
count_birth_order2/3 -0.00554 0.004017 -1.379 -0.01341 0.002334 fixed
count_birth_order3/3 0.001302 0.004497 0.2896 -0.007511 0.01012 fixed
count_birth_order1/4 -0.001762 0.004341 -0.406 -0.01027 0.006746 fixed
count_birth_order2/4 -0.001991 0.004433 -0.4491 -0.01068 0.006698 fixed
count_birth_order3/4 0.0011 0.004877 0.2256 -0.008458 0.01066 fixed
count_birth_order4/4 -0.003243 0.005046 -0.6427 -0.01313 0.006647 fixed
count_birth_order1/5 -0.003534 0.005172 -0.6833 -0.01367 0.006603 fixed
count_birth_order2/5 -0.01611 0.005572 -2.891 -0.02703 -0.00519 fixed
count_birth_order3/5 0.0009379 0.005428 0.1728 -0.009702 0.01158 fixed
count_birth_order4/5 0.001 0.005696 0.1756 -0.01016 0.01216 fixed
count_birth_order5/5 0.001112 0.005673 0.196 -0.01001 0.01223 fixed
count_birth_order1/5+ 0.001216 0.00493 0.2466 -0.008447 0.01088 fixed
count_birth_order2/5+ -0.008076 0.005152 -1.568 -0.01817 0.002021 fixed
count_birth_order3/5+ -0.01154 0.005022 -2.297 -0.02138 -0.001693 fixed
count_birth_order4/5+ 0.0009338 0.004821 0.1937 -0.008515 0.01038 fixed
count_birth_order5/5+ -0.003345 0.00499 -0.6702 -0.01313 0.006436 fixed
count_birth_order5+/5+ -0.009414 0.003609 -2.609 -0.01649 -0.002341 fixed
sd_(Intercept).mother_pidlink 0 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.06528 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -16498 -16423 8260 -16520 NA NA NA
12 -16497 -16416 8261 -16521 1.66 1 0.1976
16 -16499 -16391 8266 -16531 9.831 4 0.04338
26 -16491 -16315 8271 -16543 11.8 10 0.2989

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.018 0.02859 35.61 0.9619 1.074 fixed
poly(age, 3, raw = TRUE)1 -0.00255 0.003245 -0.7858 -0.00891 0.00381 fixed
poly(age, 3, raw = TRUE)2 0.0001017 0.0001159 0.8777 -0.0001254 0.0003288 fixed
poly(age, 3, raw = TRUE)3 -0.00000123 0.000001314 -0.936 -0.000003805 0.000001345 fixed
male -0.00009463 0.001664 -0.05687 -0.003356 0.003167 fixed
sibling_count3 -0.001208 0.00246 -0.4911 -0.006031 0.003614 fixed
sibling_count4 -0.003839 0.002643 -1.452 -0.00902 0.001341 fixed
sibling_count5 -0.003218 0.003094 -1.04 -0.009281 0.002845 fixed
sibling_count5+ -0.006981 0.002649 -2.636 -0.01217 -0.00179 fixed
sd_(Intercept).mother_pidlink 0.00000001695 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.06502 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.019 0.02859 35.62 0.9625 1.075 fixed
birth_order -0.0006737 0.0005522 -1.22 -0.001756 0.0004086 fixed
poly(age, 3, raw = TRUE)1 -0.002525 0.003245 -0.778 -0.008884 0.003835 fixed
poly(age, 3, raw = TRUE)2 0.0001021 0.0001159 0.8812 -0.000125 0.0003292 fixed
poly(age, 3, raw = TRUE)3 -0.00000126 0.000001314 -0.9586 -0.000003835 0.000001316 fixed
male -0.0000827 0.001664 -0.0497 -0.003344 0.003178 fixed
sibling_count3 -0.0008808 0.002475 -0.3559 -0.005732 0.00397 fixed
sibling_count4 -0.0031 0.002712 -1.143 -0.008415 0.002215 fixed
sibling_count5 -0.002043 0.00324 -0.6306 -0.008393 0.004307 fixed
sibling_count5+ -0.004499 0.00334 -1.347 -0.01104 0.002047 fixed
sd_(Intercept).mother_pidlink 0.00000001695 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.06501 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.021 0.02862 35.68 0.9653 1.077 fixed
poly(age, 3, raw = TRUE)1 -0.002776 0.003246 -0.8553 -0.009138 0.003585 fixed
poly(age, 3, raw = TRUE)2 0.0001105 0.0001159 0.9533 -0.0001167 0.0003377 fixed
poly(age, 3, raw = TRUE)3 -0.000001342 0.000001315 -1.021 -0.000003918 0.000001235 fixed
male -0.0001589 0.001664 -0.0955 -0.00342 0.003102 fixed
sibling_count3 -0.001365 0.002536 -0.5382 -0.006336 0.003606 fixed
sibling_count4 -0.004754 0.002838 -1.675 -0.01032 0.0008094 fixed
sibling_count5 -0.003792 0.003416 -1.11 -0.01049 0.002903 fixed
sibling_count5+ -0.006122 0.003464 -1.768 -0.01291 0.0006664 fixed
birth_order_nonlinear2 -0.004324 0.002171 -1.992 -0.008579 -0.0000701 fixed
birth_order_nonlinear3 0.0001426 0.002671 0.05338 -0.005093 0.005378 fixed
birth_order_nonlinear4 0.00281 0.003394 0.828 -0.003842 0.009463 fixed
birth_order_nonlinear5 -0.004068 0.004315 -0.9426 -0.01253 0.00439 fixed
birth_order_nonlinear5+ -0.004234 0.004176 -1.014 -0.01242 0.003951 fixed
sd_(Intercept).mother_pidlink 0 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.065 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.021 0.02867 35.61 0.9647 1.077 fixed
poly(age, 3, raw = TRUE)1 -0.002721 0.003251 -0.8371 -0.009092 0.00365 fixed
poly(age, 3, raw = TRUE)2 0.0001088 0.0001161 0.9369 -0.0001188 0.0003363 fixed
poly(age, 3, raw = TRUE)3 -0.000001325 0.000001317 -1.006 -0.000003906 0.000001257 fixed
male -0.00008317 0.001664 -0.04997 -0.003345 0.003179 fixed
count_birth_order2/2 -0.004503 0.003796 -1.186 -0.01194 0.002937 fixed
count_birth_order1/3 -0.001611 0.003351 -0.4808 -0.00818 0.004957 fixed
count_birth_order2/3 -0.00631 0.0037 -1.705 -0.01356 0.0009425 fixed
count_birth_order3/3 -0.0002261 0.004069 -0.05556 -0.0082 0.007748 fixed
count_birth_order1/4 -0.008884 0.004156 -2.137 -0.01703 -0.0007379 fixed
count_birth_order2/4 -0.006667 0.004294 -1.552 -0.01508 0.00175 fixed
count_birth_order3/4 -0.0003684 0.004503 -0.08182 -0.009195 0.008458 fixed
count_birth_order4/4 -0.004377 0.004758 -0.9199 -0.0137 0.004948 fixed
count_birth_order1/5 -0.0004467 0.005644 -0.07915 -0.01151 0.01061 fixed
count_birth_order2/5 -0.01668 0.006279 -2.656 -0.02898 -0.00437 fixed
count_birth_order3/5 -0.0006599 0.005896 -0.1119 -0.01222 0.0109 fixed
count_birth_order4/5 -0.000488 0.0058 -0.08414 -0.01186 0.01088 fixed
count_birth_order5/5 -0.008039 0.006139 -1.309 -0.02007 0.003993 fixed
count_birth_order1/5+ -0.0002868 0.005746 -0.04992 -0.01155 0.01097 fixed
count_birth_order2/5+ -0.006855 0.005725 -1.197 -0.01808 0.004366 fixed
count_birth_order3/5+ -0.01792 0.005558 -3.225 -0.02882 -0.007029 fixed
count_birth_order4/5+ -0.0006839 0.005444 -0.1256 -0.01135 0.009987 fixed
count_birth_order5/5+ -0.0102 0.005072 -2.011 -0.02014 -0.0002581 fixed
count_birth_order5+/5+ -0.01043 0.003711 -2.812 -0.01771 -0.00316 fixed
sd_(Intercept).mother_pidlink 0.000000008137 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.06498 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -16083 -16009 8052 -16105 NA NA NA
12 -16082 -16002 8053 -16106 1.491 1 0.2221
16 -16081 -15974 8057 -16113 6.603 4 0.1584
26 -16075 -15901 8064 -16127 14.38 10 0.1565

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Work

Income Last Month (log) - z-standardized

birthorder <- birthorder %>% mutate(outcome = wage_last_month_z)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.166 0.3787 -5.72 -2.909 -1.424 fixed
poly(age, 3, raw = TRUE)1 0.1631 0.03571 4.567 0.09308 0.233 fixed
poly(age, 3, raw = TRUE)2 -0.003976 0.001063 -3.739 -0.00606 -0.001892 fixed
poly(age, 3, raw = TRUE)3 0.00003055 0.00001006 3.036 0.00001083 0.00005027 fixed
male 0.1388 0.02631 5.276 0.08727 0.1904 fixed
sibling_count3 0.1019 0.05214 1.954 -0.0003369 0.204 fixed
sibling_count4 0.05405 0.05208 1.038 -0.04802 0.1561 fixed
sibling_count5 0.05158 0.05423 0.9512 -0.0547 0.1579 fixed
sibling_count5+ -0.003651 0.04276 -0.08539 -0.08745 0.08015 fixed
sd_(Intercept).mother_pidlink 0.2708 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.95 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.15 0.3789 -5.675 -2.892 -1.407 fixed
birth_order 0.007971 0.005175 1.54 -0.002171 0.01811 fixed
poly(age, 3, raw = TRUE)1 0.1595 0.03578 4.457 0.08934 0.2296 fixed
poly(age, 3, raw = TRUE)2 -0.003854 0.001066 -3.615 -0.005943 -0.001764 fixed
poly(age, 3, raw = TRUE)3 0.0000294 0.00001009 2.914 0.000009629 0.00004917 fixed
male 0.1384 0.02631 5.26 0.08684 0.19 fixed
sibling_count3 0.0995 0.05217 1.907 -0.002746 0.2017 fixed
sibling_count4 0.04798 0.05223 0.9185 -0.0544 0.1504 fixed
sibling_count5 0.04152 0.05463 0.7601 -0.06554 0.1486 fixed
sibling_count5+ -0.03363 0.04698 -0.7158 -0.1257 0.05845 fixed
sd_(Intercept).mother_pidlink 0.2729 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9493 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.148 0.3795 -5.66 -2.892 -1.404 fixed
poly(age, 3, raw = TRUE)1 0.1602 0.03579 4.475 0.09002 0.2303 fixed
poly(age, 3, raw = TRUE)2 -0.003894 0.001067 -3.65 -0.005985 -0.001803 fixed
poly(age, 3, raw = TRUE)3 0.00002995 0.0000101 2.965 0.00001015 0.00004975 fixed
male 0.1388 0.02631 5.274 0.08721 0.1904 fixed
sibling_count3 0.09578 0.05293 1.809 -0.007966 0.1995 fixed
sibling_count4 0.03642 0.05393 0.6753 -0.06928 0.1421 fixed
sibling_count5 0.01732 0.05692 0.3043 -0.09424 0.1289 fixed
sibling_count5+ -0.04504 0.0495 -0.9099 -0.1421 0.05198 fixed
birth_order_nonlinear2 0.02294 0.03805 0.6028 -0.05164 0.09751 fixed
birth_order_nonlinear3 0.03579 0.04407 0.8122 -0.05058 0.1222 fixed
birth_order_nonlinear4 0.07121 0.04928 1.445 -0.02536 0.1678 fixed
birth_order_nonlinear5 0.1192 0.05594 2.131 0.009576 0.2288 fixed
birth_order_nonlinear5+ 0.0531 0.04599 1.155 -0.03705 0.1433 fixed
sd_(Intercept).mother_pidlink 0.2737 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9492 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.12 0.3809 -5.565 -2.866 -1.373 fixed
poly(age, 3, raw = TRUE)1 0.1581 0.03585 4.409 0.0878 0.2283 fixed
poly(age, 3, raw = TRUE)2 -0.00382 0.001069 -3.574 -0.005915 -0.001725 fixed
poly(age, 3, raw = TRUE)3 0.00002915 0.00001012 2.88 0.000009315 0.00004899 fixed
male 0.1379 0.02635 5.235 0.0863 0.1896 fixed
count_birth_order2/2 -0.006871 0.07714 -0.08907 -0.1581 0.1443 fixed
count_birth_order1/3 0.1024 0.06942 1.475 -0.03367 0.2384 fixed
count_birth_order2/3 0.08335 0.0791 1.054 -0.07168 0.2384 fixed
count_birth_order3/3 0.1205 0.08859 1.361 -0.05309 0.2942 fixed
count_birth_order1/4 0.06485 0.07708 0.8413 -0.08622 0.2159 fixed
count_birth_order2/4 0.08698 0.08249 1.054 -0.0747 0.2487 fixed
count_birth_order3/4 -0.04521 0.08657 -0.5222 -0.2149 0.1245 fixed
count_birth_order4/4 0.108 0.09454 1.142 -0.07732 0.2933 fixed
count_birth_order1/5 -0.03409 0.08715 -0.3912 -0.2049 0.1367 fixed
count_birth_order2/5 0.1125 0.09623 1.169 -0.07613 0.3011 fixed
count_birth_order3/5 0.05178 0.09796 0.5285 -0.1402 0.2438 fixed
count_birth_order4/5 0.03552 0.09973 0.3561 -0.1599 0.231 fixed
count_birth_order5/5 0.1267 0.1031 1.229 -0.07539 0.3288 fixed
count_birth_order1/5+ -0.09157 0.06947 -1.318 -0.2277 0.04459 fixed
count_birth_order2/5+ -0.06085 0.07256 -0.8386 -0.2031 0.08136 fixed
count_birth_order3/5+ 0.0349 0.07161 0.4874 -0.1055 0.1753 fixed
count_birth_order4/5+ 0.02553 0.06944 0.3677 -0.1106 0.1616 fixed
count_birth_order5/5+ 0.06345 0.07022 0.9035 -0.07419 0.2011 fixed
count_birth_order5+/5+ -0.002149 0.05532 -0.03884 -0.1106 0.1063 fixed
sd_(Intercept).mother_pidlink 0.2764 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9487 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 16856 16930 -8417 16834 NA NA NA
12 16856 16936 -8416 16832 2.37 1 0.1237
16 16861 16968 -8415 16829 2.788 4 0.5939
26 16875 17049 -8411 16823 6.743 10 0.7495

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.967 0.8516 -3.484 -4.636 -1.298 fixed
poly(age, 3, raw = TRUE)1 0.24 0.09033 2.656 0.06292 0.417 fixed
poly(age, 3, raw = TRUE)2 -0.00612 0.003071 -1.993 -0.01214 -0.0001006 fixed
poly(age, 3, raw = TRUE)3 0.00005134 0.00003351 1.532 -0.00001433 0.000117 fixed
male 0.1734 0.03918 4.426 0.09659 0.2502 fixed
sibling_count3 0.05539 0.06069 0.9127 -0.06356 0.1743 fixed
sibling_count4 -0.05519 0.06241 -0.8843 -0.1775 0.06713 fixed
sibling_count5 -0.1223 0.07112 -1.72 -0.2617 0.01709 fixed
sibling_count5+ -0.1055 0.06141 -1.718 -0.2258 0.01488 fixed
sd_(Intercept).mother_pidlink 0.1431 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9554 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.949 0.851 -3.465 -4.617 -1.281 fixed
birth_order 0.02897 0.01228 2.359 0.004905 0.05304 fixed
poly(age, 3, raw = TRUE)1 0.2348 0.09028 2.601 0.05783 0.4117 fixed
poly(age, 3, raw = TRUE)2 -0.005999 0.003069 -1.955 -0.01201 0.00001559 fixed
poly(age, 3, raw = TRUE)3 0.00005109 0.00003348 1.526 -0.00001453 0.0001167 fixed
male 0.1711 0.03915 4.371 0.09441 0.2479 fixed
sibling_count3 0.04128 0.06096 0.6772 -0.07819 0.1608 fixed
sibling_count4 -0.08726 0.06386 -1.367 -0.2124 0.03789 fixed
sibling_count5 -0.1757 0.07463 -2.355 -0.322 -0.02945 fixed
sibling_count5+ -0.2097 0.07565 -2.772 -0.3579 -0.0614 fixed
sd_(Intercept).mother_pidlink 0.1526 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9531 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.933 0.8519 -3.443 -4.603 -1.264 fixed
poly(age, 3, raw = TRUE)1 0.2365 0.09037 2.617 0.05938 0.4136 fixed
poly(age, 3, raw = TRUE)2 -0.006059 0.003072 -1.972 -0.01208 -0.00003838 fixed
poly(age, 3, raw = TRUE)3 0.0000517 0.00003351 1.543 -0.00001399 0.0001174 fixed
male 0.1724 0.03916 4.402 0.09565 0.2492 fixed
sibling_count3 0.02577 0.06226 0.4139 -0.09625 0.1478 fixed
sibling_count4 -0.1117 0.06659 -1.678 -0.2422 0.01879 fixed
sibling_count5 -0.1958 0.07894 -2.48 -0.3505 -0.04107 fixed
sibling_count5+ -0.2197 0.07809 -2.813 -0.3727 -0.06662 fixed
birth_order_nonlinear2 0.02232 0.05137 0.4345 -0.07837 0.123 fixed
birth_order_nonlinear3 0.1267 0.06161 2.057 0.005969 0.2475 fixed
birth_order_nonlinear4 0.1321 0.07471 1.769 -0.01429 0.2786 fixed
birth_order_nonlinear5 0.0958 0.09391 1.02 -0.08827 0.2799 fixed
birth_order_nonlinear5+ 0.1847 0.09213 2.004 0.004083 0.3652 fixed
sd_(Intercept).mother_pidlink 0.1585 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9525 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.924 0.8537 -3.424 -4.597 -1.25 fixed
poly(age, 3, raw = TRUE)1 0.2359 0.09068 2.601 0.05814 0.4136 fixed
poly(age, 3, raw = TRUE)2 -0.006003 0.003083 -1.947 -0.01205 0.00003935 fixed
poly(age, 3, raw = TRUE)3 0.00005074 0.00003364 1.508 -0.00001519 0.0001167 fixed
male 0.1682 0.03928 4.281 0.09118 0.2451 fixed
count_birth_order2/2 -0.01186 0.09694 -0.1223 -0.2019 0.1781 fixed
count_birth_order1/3 0.002427 0.08014 0.03029 -0.1546 0.1595 fixed
count_birth_order2/3 0.08795 0.09083 0.9683 -0.09007 0.266 fixed
count_birth_order3/3 0.09969 0.1004 0.9929 -0.09709 0.2965 fixed
count_birth_order1/4 -0.193 0.09389 -2.056 -0.3771 -0.009012 fixed
count_birth_order2/4 -0.03204 0.09714 -0.3299 -0.2224 0.1584 fixed
count_birth_order3/4 0.05193 0.1026 0.5059 -0.1492 0.2531 fixed
count_birth_order4/4 -0.03512 0.11 -0.3193 -0.2507 0.1805 fixed
count_birth_order1/5 -0.0734 0.127 -0.5779 -0.3223 0.1755 fixed
count_birth_order2/5 -0.2006 0.1483 -1.353 -0.4913 0.09001 fixed
count_birth_order3/5 -0.1334 0.1329 -1.003 -0.3939 0.1272 fixed
count_birth_order4/5 -0.06657 0.1264 -0.5265 -0.3144 0.1812 fixed
count_birth_order5/5 -0.2006 0.1302 -1.54 -0.4559 0.05465 fixed
count_birth_order1/5+ -0.204 0.1252 -1.629 -0.4494 0.0414 fixed
count_birth_order2/5+ -0.4087 0.1263 -3.236 -0.6563 -0.1611 fixed
count_birth_order3/5+ -0.06434 0.1266 -0.5081 -0.3125 0.1839 fixed
count_birth_order4/5+ -0.05434 0.1165 -0.4663 -0.2828 0.1741 fixed
count_birth_order5/5+ -0.07096 0.1149 -0.6177 -0.2961 0.1542 fixed
count_birth_order5+/5+ -0.04591 0.08527 -0.5384 -0.213 0.1212 fixed
sd_(Intercept).mother_pidlink 0.1652 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9515 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 7048 7112 -3513 7026 NA NA NA
12 7045 7115 -3510 7021 5.566 1 0.01831
16 7051 7144 -3509 7019 1.971 4 0.741
26 7061 7213 -3505 7009 9.311 10 0.5029

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.021 0.8508 -3.551 -4.689 -1.354 fixed
poly(age, 3, raw = TRUE)1 0.2478 0.09033 2.744 0.07078 0.4249 fixed
poly(age, 3, raw = TRUE)2 -0.006428 0.003071 -2.093 -0.01245 -0.0004093 fixed
poly(age, 3, raw = TRUE)3 0.00005481 0.00003351 1.636 -0.00001086 0.0001205 fixed
male 0.1666 0.03915 4.256 0.08989 0.2433 fixed
sibling_count3 0.03445 0.06704 0.5139 -0.09694 0.1658 fixed
sibling_count4 0.00218 0.06772 0.03219 -0.1305 0.1349 fixed
sibling_count5 -0.1056 0.0713 -1.481 -0.2454 0.03414 fixed
sibling_count5+ -0.08992 0.06331 -1.42 -0.214 0.03417 fixed
sd_(Intercept).mother_pidlink 0.1499 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9564 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.016 0.8506 -3.546 -4.683 -1.349 fixed
birth_order 0.01776 0.01074 1.653 -0.003294 0.03882 fixed
poly(age, 3, raw = TRUE)1 0.2451 0.09032 2.714 0.06808 0.4221 fixed
poly(age, 3, raw = TRUE)2 -0.006365 0.00307 -2.073 -0.01238 -0.0003467 fixed
poly(age, 3, raw = TRUE)3 0.00005472 0.0000335 1.633 -0.00001094 0.0001204 fixed
male 0.1651 0.03914 4.218 0.08839 0.2418 fixed
sibling_count3 0.0256 0.06726 0.3806 -0.1062 0.1574 fixed
sibling_count4 -0.01668 0.06869 -0.2428 -0.1513 0.118 fixed
sibling_count5 -0.1356 0.07361 -1.842 -0.2799 0.00866 fixed
sibling_count5+ -0.1524 0.07377 -2.066 -0.297 -0.007788 fixed
sd_(Intercept).mother_pidlink 0.1605 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9544 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.025 0.8511 -3.555 -4.693 -1.357 fixed
poly(age, 3, raw = TRUE)1 0.247 0.09037 2.733 0.06989 0.4241 fixed
poly(age, 3, raw = TRUE)2 -0.006441 0.003072 -2.097 -0.01246 -0.00042 fixed
poly(age, 3, raw = TRUE)3 0.00005564 0.00003352 1.66 -0.00001005 0.0001213 fixed
male 0.1661 0.03914 4.244 0.08938 0.2428 fixed
sibling_count3 0.005788 0.06854 0.08445 -0.1285 0.1401 fixed
sibling_count4 -0.05279 0.07109 -0.7426 -0.1921 0.08654 fixed
sibling_count5 -0.167 0.07725 -2.161 -0.3184 -0.01555 fixed
sibling_count5+ -0.1714 0.07608 -2.254 -0.3206 -0.02234 fixed
birth_order_nonlinear2 0.05834 0.0521 1.12 -0.04379 0.1605 fixed
birth_order_nonlinear3 0.1224 0.06201 1.974 0.0008591 0.2439 fixed
birth_order_nonlinear4 0.1577 0.07258 2.173 0.01544 0.2999 fixed
birth_order_nonlinear5 0.07475 0.08919 0.8381 -0.1001 0.2496 fixed
birth_order_nonlinear5+ 0.1222 0.08292 1.473 -0.04038 0.2847 fixed
sd_(Intercept).mother_pidlink 0.1634 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9539 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.005 0.8535 -3.521 -4.678 -1.333 fixed
poly(age, 3, raw = TRUE)1 0.2466 0.09072 2.718 0.06875 0.4244 fixed
poly(age, 3, raw = TRUE)2 -0.006401 0.003084 -2.075 -0.01245 -0.0003561 fixed
poly(age, 3, raw = TRUE)3 0.00005498 0.00003365 1.634 -0.00001098 0.0001209 fixed
male 0.1643 0.03928 4.183 0.08731 0.2413 fixed
count_birth_order2/2 -0.01229 0.1097 -0.112 -0.2272 0.2027 fixed
count_birth_order1/3 -0.03131 0.08855 -0.3535 -0.2049 0.1422 fixed
count_birth_order2/3 0.1016 0.09821 1.035 -0.09088 0.2941 fixed
count_birth_order3/3 0.054 0.1103 0.4897 -0.1621 0.2702 fixed
count_birth_order1/4 -0.1062 0.09942 -1.068 -0.301 0.0887 fixed
count_birth_order2/4 0.06707 0.1002 0.6692 -0.1294 0.2635 fixed
count_birth_order3/4 0.02235 0.1126 0.1984 -0.1984 0.2431 fixed
count_birth_order4/4 0.02293 0.1206 0.19 -0.2135 0.2594 fixed
count_birth_order1/5 -0.1291 0.116 -1.113 -0.3564 0.09827 fixed
count_birth_order2/5 -0.1532 0.1278 -1.199 -0.4036 0.09726 fixed
count_birth_order3/5 -0.05305 0.1277 -0.4155 -0.3033 0.1972 fixed
count_birth_order4/5 -0.03084 0.1287 -0.2395 -0.2832 0.2215 fixed
count_birth_order5/5 -0.1853 0.1282 -1.446 -0.4366 0.06595 fixed
count_birth_order1/5+ -0.2191 0.1105 -1.984 -0.4356 -0.002651 fixed
count_birth_order2/5+ -0.2885 0.1175 -2.456 -0.5188 -0.05829 fixed
count_birth_order3/5+ 0.006948 0.1182 0.05879 -0.2247 0.2386 fixed
count_birth_order4/5+ 0.008072 0.1093 0.07383 -0.2062 0.2223 fixed
count_birth_order5/5+ -0.06426 0.1169 -0.5495 -0.2935 0.165 fixed
count_birth_order5+/5+ -0.07115 0.08545 -0.8326 -0.2386 0.09633 fixed
sd_(Intercept).mother_pidlink 0.1732 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9527 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 7103 7167 -3541 7081 NA NA NA
12 7102 7173 -3539 7078 2.723 1 0.0989
16 7106 7200 -3537 7074 3.99 4 0.4074
26 7119 7271 -3533 7067 7.706 10 0.6575

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.95 0.8647 -3.411 -4.644 -1.255 fixed
poly(age, 3, raw = TRUE)1 0.2362 0.09172 2.575 0.05644 0.416 fixed
poly(age, 3, raw = TRUE)2 -0.005959 0.003116 -1.912 -0.01207 0.000149 fixed
poly(age, 3, raw = TRUE)3 0.00004927 0.00003398 1.45 -0.00001733 0.0001159 fixed
male 0.1752 0.03992 4.388 0.09692 0.2534 fixed
sibling_count3 0.04336 0.06038 0.7181 -0.07499 0.1617 fixed
sibling_count4 -0.08107 0.06277 -1.292 -0.2041 0.04196 fixed
sibling_count5 -0.09708 0.07355 -1.32 -0.2412 0.04707 fixed
sibling_count5+ -0.08266 0.06241 -1.324 -0.205 0.03966 fixed
sd_(Intercept).mother_pidlink 0.15 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9632 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.931 0.8643 -3.391 -4.625 -1.237 fixed
birth_order 0.02467 0.01273 1.938 -0.0002824 0.04963 fixed
poly(age, 3, raw = TRUE)1 0.2313 0.09171 2.523 0.0516 0.4111 fixed
poly(age, 3, raw = TRUE)2 -0.005839 0.003115 -1.874 -0.01195 0.0002667 fixed
poly(age, 3, raw = TRUE)3 0.00004885 0.00003396 1.438 -0.00001772 0.0001154 fixed
male 0.1739 0.03991 4.357 0.09568 0.2521 fixed
sibling_count3 0.03093 0.06071 0.5095 -0.08806 0.1499 fixed
sibling_count4 -0.108 0.06429 -1.68 -0.234 0.01797 fixed
sibling_count5 -0.1412 0.07701 -1.834 -0.2921 0.009731 fixed
sibling_count5+ -0.1711 0.07736 -2.212 -0.3228 -0.01952 fixed
sd_(Intercept).mother_pidlink 0.1571 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9616 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.945 0.8651 -3.404 -4.641 -1.25 fixed
poly(age, 3, raw = TRUE)1 0.2361 0.09179 2.573 0.05623 0.416 fixed
poly(age, 3, raw = TRUE)2 -0.006009 0.003118 -1.927 -0.01212 0.0001024 fixed
poly(age, 3, raw = TRUE)3 0.0000507 0.000034 1.491 -0.00001593 0.0001173 fixed
male 0.1757 0.03993 4.402 0.09749 0.254 fixed
sibling_count3 0.009872 0.06207 0.159 -0.1118 0.1315 fixed
sibling_count4 -0.1259 0.06698 -1.88 -0.2572 0.005376 fixed
sibling_count5 -0.1521 0.08111 -1.875 -0.3111 0.006863 fixed
sibling_count5+ -0.1838 0.07992 -2.3 -0.3404 -0.02714 fixed
birth_order_nonlinear2 0.01217 0.05179 0.235 -0.08934 0.1137 fixed
birth_order_nonlinear3 0.1384 0.0622 2.225 0.01651 0.2603 fixed
birth_order_nonlinear4 0.05726 0.07794 0.7346 -0.0955 0.21 fixed
birth_order_nonlinear5 0.07804 0.09835 0.7935 -0.1147 0.2708 fixed
birth_order_nonlinear5+ 0.1773 0.09559 1.855 -0.01002 0.3647 fixed
sd_(Intercept).mother_pidlink 0.1621 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9608 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.969 0.8669 -3.426 -4.668 -1.27 fixed
poly(age, 3, raw = TRUE)1 0.2408 0.09211 2.615 0.06029 0.4214 fixed
poly(age, 3, raw = TRUE)2 -0.006135 0.003129 -1.96 -0.01227 -0.000001385 fixed
poly(age, 3, raw = TRUE)3 0.00005172 0.00003412 1.516 -0.00001516 0.0001186 fixed
male 0.1726 0.04006 4.308 0.09407 0.2511 fixed
count_birth_order2/2 -0.07046 0.09561 -0.7369 -0.2579 0.1169 fixed
count_birth_order1/3 -0.04347 0.08004 -0.5431 -0.2003 0.1134 fixed
count_birth_order2/3 0.06615 0.09109 0.7262 -0.1124 0.2447 fixed
count_birth_order3/3 0.08534 0.09871 0.8646 -0.1081 0.2788 fixed
count_birth_order1/4 -0.2178 0.09523 -2.287 -0.4044 -0.03112 fixed
count_birth_order2/4 -0.0797 0.09775 -0.8153 -0.2713 0.1119 fixed
count_birth_order3/4 0.04818 0.1035 0.4657 -0.1546 0.251 fixed
count_birth_order4/4 -0.1601 0.1133 -1.414 -0.3821 0.06185 fixed
count_birth_order1/5 -0.08318 0.1309 -0.6352 -0.3398 0.1735 fixed
count_birth_order2/5 -0.1651 0.1522 -1.085 -0.4634 0.1332 fixed
count_birth_order3/5 -0.09123 0.1405 -0.6492 -0.3667 0.1842 fixed
count_birth_order4/5 -0.1192 0.1345 -0.8861 -0.3827 0.1444 fixed
count_birth_order5/5 -0.1676 0.1392 -1.204 -0.4405 0.1053 fixed
count_birth_order1/5+ -0.1719 0.1281 -1.342 -0.4229 0.07922 fixed
count_birth_order2/5+ -0.3917 0.1323 -2.961 -0.6509 -0.1324 fixed
count_birth_order3/5+ -0.06893 0.1285 -0.5364 -0.3208 0.1829 fixed
count_birth_order4/5+ -0.07846 0.1223 -0.6416 -0.3181 0.1612 fixed
count_birth_order5/5+ -0.08894 0.1182 -0.7525 -0.3206 0.1427 fixed
count_birth_order5+/5+ -0.03323 0.08726 -0.3809 -0.2042 0.1378 fixed
sd_(Intercept).mother_pidlink 0.1651 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9606 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 6958 7022 -3468 6936 NA NA NA
12 6956 7026 -3466 6932 3.756 1 0.05262
16 6960 7053 -3464 6928 3.587 4 0.4648
26 6971 7123 -3460 6919 8.955 10 0.5364

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Income last year (log) - z-standardized

birthorder <- birthorder %>% mutate(outcome = wage_last_year_z)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.932 0.3808 -10.33 -4.679 -3.186 fixed
poly(age, 3, raw = TRUE)1 0.3021 0.03594 8.408 0.2317 0.3726 fixed
poly(age, 3, raw = TRUE)2 -0.007364 0.001071 -6.873 -0.009464 -0.005264 fixed
poly(age, 3, raw = TRUE)3 0.00005769 0.00001015 5.683 0.00003779 0.00007758 fixed
male 0.1267 0.02642 4.798 0.07497 0.1785 fixed
sibling_count3 0.0766 0.05286 1.449 -0.02701 0.1802 fixed
sibling_count4 0.04296 0.05286 0.8128 -0.06064 0.1466 fixed
sibling_count5 0.03471 0.05503 0.6307 -0.07315 0.1426 fixed
sibling_count5+ -0.04989 0.04336 -1.151 -0.1349 0.03509 fixed
sd_(Intercept).mother_pidlink 0.341 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9299 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.925 0.3809 -10.3 -4.672 -3.179 fixed
birth_order 0.003522 0.005228 0.6737 -0.006725 0.01377 fixed
poly(age, 3, raw = TRUE)1 0.3005 0.03602 8.344 0.2299 0.3711 fixed
poly(age, 3, raw = TRUE)2 -0.007309 0.001075 -6.801 -0.009415 -0.005202 fixed
poly(age, 3, raw = TRUE)3 0.00005716 0.00001018 5.614 0.00003721 0.00007712 fixed
male 0.1265 0.02642 4.789 0.07475 0.1783 fixed
sibling_count3 0.07559 0.05289 1.429 -0.02807 0.1793 fixed
sibling_count4 0.0403 0.05301 0.7601 -0.06361 0.1442 fixed
sibling_count5 0.03033 0.05542 0.5472 -0.07829 0.139 fixed
sibling_count5+ -0.06306 0.04756 -1.326 -0.1563 0.03016 fixed
sd_(Intercept).mother_pidlink 0.3417 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9298 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.911 0.3816 -10.25 -4.659 -3.163 fixed
poly(age, 3, raw = TRUE)1 0.3 0.03603 8.327 0.2294 0.3707 fixed
poly(age, 3, raw = TRUE)2 -0.007284 0.001075 -6.773 -0.009392 -0.005176 fixed
poly(age, 3, raw = TRUE)3 0.00005682 0.0000102 5.573 0.00003684 0.0000768 fixed
male 0.1269 0.02642 4.801 0.07508 0.1787 fixed
sibling_count3 0.08191 0.05362 1.528 -0.02318 0.187 fixed
sibling_count4 0.04998 0.05468 0.9141 -0.05719 0.1572 fixed
sibling_count5 0.02631 0.05763 0.4565 -0.08664 0.1393 fixed
sibling_count5+ -0.06185 0.05001 -1.237 -0.1599 0.03617 fixed
birth_order_nonlinear2 -0.0194 0.03805 -0.5098 -0.09398 0.05518 fixed
birth_order_nonlinear3 -0.0268 0.04408 -0.608 -0.1132 0.0596 fixed
birth_order_nonlinear4 -0.01244 0.04932 -0.2522 -0.1091 0.08423 fixed
birth_order_nonlinear5 0.07847 0.05601 1.401 -0.03131 0.1883 fixed
birth_order_nonlinear5+ 0.006064 0.04625 0.1311 -0.08458 0.09671 fixed
sd_(Intercept).mother_pidlink 0.3407 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9302 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -3.899 0.3829 -10.18 -4.649 -3.148 fixed
poly(age, 3, raw = TRUE)1 0.2982 0.03608 8.265 0.2275 0.3689 fixed
poly(age, 3, raw = TRUE)2 -0.007207 0.001077 -6.691 -0.009319 -0.005096 fixed
poly(age, 3, raw = TRUE)3 0.00005593 0.00001021 5.476 0.00003591 0.00007595 fixed
male 0.1262 0.02645 4.771 0.07434 0.178 fixed
count_birth_order2/2 -0.02015 0.07713 -0.2613 -0.1713 0.131 fixed
count_birth_order1/3 0.09485 0.06972 1.36 -0.04181 0.2315 fixed
count_birth_order2/3 0.03908 0.0799 0.4891 -0.1175 0.1957 fixed
count_birth_order3/3 0.05765 0.08899 0.6479 -0.1168 0.2321 fixed
count_birth_order1/4 0.1217 0.07784 1.563 -0.03087 0.2743 fixed
count_birth_order2/4 0.02243 0.08287 0.2707 -0.14 0.1849 fixed
count_birth_order3/4 -0.07628 0.08696 -0.8772 -0.2467 0.09416 fixed
count_birth_order4/4 0.04495 0.09473 0.4745 -0.1407 0.2306 fixed
count_birth_order1/5 0.07361 0.08756 0.8407 -0.09801 0.2452 fixed
count_birth_order2/5 0.0388 0.09613 0.4036 -0.1496 0.2272 fixed
count_birth_order3/5 0.03883 0.09843 0.3945 -0.1541 0.2318 fixed
count_birth_order4/5 -0.1091 0.1005 -1.086 -0.3061 0.08788 fixed
count_birth_order5/5 0.07412 0.1036 0.7154 -0.129 0.2772 fixed
count_birth_order1/5+ -0.1477 0.06989 -2.113 -0.2846 -0.01067 fixed
count_birth_order2/5+ -0.07679 0.07276 -1.055 -0.2194 0.06581 fixed
count_birth_order3/5+ -0.05401 0.072 -0.7501 -0.1951 0.08711 fixed
count_birth_order4/5+ -0.0383 0.0698 -0.5487 -0.1751 0.09851 fixed
count_birth_order5/5+ 0.02497 0.07063 0.3535 -0.1135 0.1634 fixed
count_birth_order5+/5+ -0.05612 0.05586 -1.005 -0.1656 0.05337 fixed
sd_(Intercept).mother_pidlink 0.3415 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9299 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 16736 16809 -8357 16714 NA NA NA
12 16737 16818 -8357 16713 0.4529 1 0.501
16 16742 16849 -8355 16710 3.397 4 0.4938
26 16752 16926 -8350 16700 10.14 10 0.4285

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -5.457 0.8887 -6.141 -7.199 -3.715 fixed
poly(age, 3, raw = TRUE)1 0.4617 0.0944 4.891 0.2767 0.6468 fixed
poly(age, 3, raw = TRUE)2 -0.01258 0.003215 -3.911 -0.01888 -0.006274 fixed
poly(age, 3, raw = TRUE)3 0.0001155 0.00003516 3.286 0.00004662 0.0001845 fixed
male 0.17 0.04053 4.196 0.0906 0.2495 fixed
sibling_count3 0.03386 0.06327 0.5352 -0.09015 0.1579 fixed
sibling_count4 -0.04681 0.06518 -0.7182 -0.1745 0.08093 fixed
sibling_count5 -0.1543 0.07425 -2.078 -0.2998 -0.008749 fixed
sibling_count5+ -0.1823 0.0642 -2.84 -0.3081 -0.05649 fixed
sd_(Intercept).mother_pidlink 0.2543 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9641 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -5.45 0.8887 -6.133 -7.192 -3.709 fixed
birth_order 0.01116 0.01275 0.8758 -0.01382 0.03615 fixed
poly(age, 3, raw = TRUE)1 0.4598 0.09444 4.869 0.2747 0.6449 fixed
poly(age, 3, raw = TRUE)2 -0.01253 0.003216 -3.897 -0.01884 -0.006228 fixed
poly(age, 3, raw = TRUE)3 0.0001155 0.00003516 3.284 0.00004655 0.0001844 fixed
male 0.1691 0.04054 4.171 0.08963 0.2485 fixed
sibling_count3 0.02843 0.06359 0.4472 -0.0962 0.1531 fixed
sibling_count4 -0.05926 0.06673 -0.8881 -0.19 0.07152 fixed
sibling_count5 -0.175 0.07795 -2.245 -0.3278 -0.02224 fixed
sibling_count5+ -0.2227 0.07907 -2.817 -0.3777 -0.06775 fixed
sd_(Intercept).mother_pidlink 0.2562 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9637 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -5.411 0.8892 -6.085 -7.154 -3.668 fixed
poly(age, 3, raw = TRUE)1 0.4567 0.09448 4.833 0.2715 0.6418 fixed
poly(age, 3, raw = TRUE)2 -0.01242 0.003218 -3.861 -0.01873 -0.006117 fixed
poly(age, 3, raw = TRUE)3 0.0001143 0.00003518 3.248 0.0000453 0.0001832 fixed
male 0.1692 0.04053 4.174 0.08973 0.2486 fixed
sibling_count3 0.01621 0.06489 0.2498 -0.111 0.1434 fixed
sibling_count4 -0.08453 0.06948 -1.216 -0.2207 0.05166 fixed
sibling_count5 -0.1747 0.08228 -2.123 -0.336 -0.01344 fixed
sibling_count5+ -0.2263 0.08151 -2.777 -0.3861 -0.06659 fixed
birth_order_nonlinear2 0.01137 0.05292 0.2149 -0.09235 0.1151 fixed
birth_order_nonlinear3 0.07625 0.06357 1.199 -0.04834 0.2008 fixed
birth_order_nonlinear4 0.101 0.07699 1.311 -0.04993 0.2519 fixed
birth_order_nonlinear5 -0.07953 0.09654 -0.8238 -0.2687 0.1097 fixed
birth_order_nonlinear5+ 0.09209 0.09551 0.9642 -0.09511 0.2793 fixed
sd_(Intercept).mother_pidlink 0.262 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.962 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -5.375 0.8908 -6.033 -7.121 -3.629 fixed
poly(age, 3, raw = TRUE)1 0.4507 0.09477 4.756 0.265 0.6365 fixed
poly(age, 3, raw = TRUE)2 -0.01218 0.003228 -3.774 -0.01851 -0.005857 fixed
poly(age, 3, raw = TRUE)3 0.0001113 0.0000353 3.153 0.00004211 0.0001805 fixed
male 0.1628 0.04062 4.007 0.08315 0.2424 fixed
count_birth_order2/2 0.05057 0.0999 0.5062 -0.1452 0.2464 fixed
count_birth_order1/3 0.008136 0.08297 0.09806 -0.1545 0.1708 fixed
count_birth_order2/3 0.11 0.09415 1.168 -0.07454 0.2945 fixed
count_birth_order3/3 0.05356 0.1036 0.5168 -0.1496 0.2567 fixed
count_birth_order1/4 -0.09397 0.09738 -0.965 -0.2848 0.09689 fixed
count_birth_order2/4 -0.06464 0.1005 -0.6428 -0.2617 0.1324 fixed
count_birth_order3/4 0.04913 0.1063 0.4623 -0.1592 0.2574 fixed
count_birth_order4/4 0.0121 0.1135 0.1066 -0.2104 0.2346 fixed
count_birth_order1/5 -0.008234 0.1311 -0.06281 -0.2652 0.2487 fixed
count_birth_order2/5 -0.1178 0.153 -0.7701 -0.4177 0.182 fixed
count_birth_order3/5 -0.1788 0.1381 -1.295 -0.4495 0.09185 fixed
count_birth_order4/5 -0.04439 0.1305 -0.3402 -0.3001 0.2113 fixed
count_birth_order5/5 -0.3668 0.1344 -2.729 -0.6302 -0.1033 fixed
count_birth_order1/5+ -0.1898 0.1294 -1.467 -0.4434 0.06372 fixed
count_birth_order2/5+ -0.4359 0.1304 -3.343 -0.6915 -0.1804 fixed
count_birth_order3/5+ -0.0447 0.1321 -0.3384 -0.3036 0.2142 fixed
count_birth_order4/5+ -0.1088 0.1208 -0.9012 -0.3455 0.1279 fixed
count_birth_order5/5+ -0.2063 0.1181 -1.747 -0.4377 0.0251 fixed
count_birth_order5+/5+ -0.122 0.0885 -1.378 -0.2954 0.05147 fixed
sd_(Intercept).mother_pidlink 0.2591 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9626 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 7164 7228 -3571 7142 NA NA NA
12 7165 7235 -3570 7141 0.7669 1 0.3812
16 7168 7261 -3568 7136 4.699 4 0.3195
26 7177 7329 -3563 7125 10.95 10 0.3616

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -5.441 0.8839 -6.156 -7.173 -3.708 fixed
poly(age, 3, raw = TRUE)1 0.4605 0.09398 4.9 0.2763 0.6447 fixed
poly(age, 3, raw = TRUE)2 -0.01255 0.003201 -3.921 -0.01883 -0.006279 fixed
poly(age, 3, raw = TRUE)3 0.0001153 0.00003501 3.292 0.00004664 0.0001839 fixed
male 0.1649 0.04032 4.089 0.08584 0.2439 fixed
sibling_count3 0.03006 0.06953 0.4324 -0.1062 0.1663 fixed
sibling_count4 0.01484 0.07034 0.2109 -0.123 0.1527 fixed
sibling_count5 -0.1237 0.07397 -1.673 -0.2687 0.02123 fixed
sibling_count5+ -0.153 0.06571 -2.328 -0.2817 -0.02416 fixed
sd_(Intercept).mother_pidlink 0.2542 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.962 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -5.441 0.884 -6.154 -7.173 -3.708 fixed
birth_order 0.001754 0.01111 0.158 -0.02002 0.02352 fixed
poly(age, 3, raw = TRUE)1 0.4603 0.09402 4.896 0.276 0.6446 fixed
poly(age, 3, raw = TRUE)2 -0.01255 0.003202 -3.919 -0.01882 -0.006272 fixed
poly(age, 3, raw = TRUE)3 0.0001153 0.00003501 3.292 0.00004663 0.0001839 fixed
male 0.1647 0.04034 4.083 0.08564 0.2438 fixed
sibling_count3 0.02918 0.06976 0.4183 -0.1076 0.1659 fixed
sibling_count4 0.01296 0.07136 0.1816 -0.1269 0.1528 fixed
sibling_count5 -0.1267 0.07636 -1.66 -0.2764 0.02293 fixed
sibling_count5+ -0.1591 0.07655 -2.079 -0.3092 -0.00911 fixed
sd_(Intercept).mother_pidlink 0.2551 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9619 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -5.442 0.8838 -6.157 -7.174 -3.71 fixed
poly(age, 3, raw = TRUE)1 0.4599 0.09399 4.893 0.2757 0.6442 fixed
poly(age, 3, raw = TRUE)2 -0.01254 0.003201 -3.916 -0.01881 -0.006262 fixed
poly(age, 3, raw = TRUE)3 0.0001151 0.000035 3.289 0.00004653 0.0001837 fixed
male 0.1643 0.0403 4.077 0.08532 0.2433 fixed
sibling_count3 0.009383 0.07102 0.1321 -0.1298 0.1486 fixed
sibling_count4 -0.0289 0.07374 -0.3919 -0.1734 0.1156 fixed
sibling_count5 -0.1455 0.07999 -1.819 -0.3023 0.01128 fixed
sibling_count5+ -0.1811 0.07882 -2.297 -0.3356 -0.0266 fixed
birth_order_nonlinear2 0.02068 0.05343 0.3871 -0.08405 0.1254 fixed
birth_order_nonlinear3 0.09024 0.06368 1.417 -0.03457 0.215 fixed
birth_order_nonlinear4 0.1398 0.07445 1.878 -0.00608 0.2858 fixed
birth_order_nonlinear5 -0.1024 0.0914 -1.121 -0.2816 0.07671 fixed
birth_order_nonlinear5+ 0.03478 0.08567 0.406 -0.1331 0.2027 fixed
sd_(Intercept).mother_pidlink 0.2586 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9602 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -5.453 0.8868 -6.149 -7.191 -3.715 fixed
poly(age, 3, raw = TRUE)1 0.4599 0.09439 4.873 0.2749 0.6449 fixed
poly(age, 3, raw = TRUE)2 -0.01251 0.003215 -3.89 -0.01881 -0.006206 fixed
poly(age, 3, raw = TRUE)3 0.0001146 0.00003516 3.259 0.00004568 0.0001835 fixed
male 0.1614 0.04046 3.989 0.08207 0.2407 fixed
count_birth_order2/2 0.03684 0.1127 0.327 -0.184 0.2577 fixed
count_birth_order1/3 -0.01654 0.09127 -0.1812 -0.1954 0.1624 fixed
count_birth_order2/3 0.1026 0.1014 1.012 -0.09611 0.3012 fixed
count_birth_order3/3 0.07168 0.1133 0.6325 -0.1505 0.2938 fixed
count_birth_order1/4 0.000209 0.1026 0.002036 -0.2009 0.2014 fixed
count_birth_order2/4 0.01402 0.1032 0.1358 -0.1883 0.2164 fixed
count_birth_order3/4 0.02956 0.1166 0.2536 -0.1989 0.258 fixed
count_birth_order4/4 0.08172 0.124 0.6593 -0.1612 0.3247 fixed
count_birth_order1/5 -0.07394 0.1193 -0.6201 -0.3077 0.1598 fixed
count_birth_order2/5 -0.1639 0.1313 -1.249 -0.4212 0.09335 fixed
count_birth_order3/5 -0.01559 0.1312 -0.1188 -0.2727 0.2415 fixed
count_birth_order4/5 0.03078 0.1323 0.2328 -0.2284 0.29 fixed
count_birth_order5/5 -0.3571 0.1317 -2.712 -0.6153 -0.09899 fixed
count_birth_order1/5+ -0.1954 0.1136 -1.72 -0.4181 0.02726 fixed
count_birth_order2/5+ -0.2732 0.1208 -2.263 -0.5099 -0.03654 fixed
count_birth_order3/5+ -0.03812 0.122 -0.3126 -0.2772 0.2009 fixed
count_birth_order4/5+ -0.03309 0.1126 -0.2937 -0.2539 0.1877 fixed
count_birth_order5/5+ -0.1922 0.1201 -1.6 -0.4277 0.04322 fixed
count_birth_order5+/5+ -0.1421 0.08825 -1.61 -0.3151 0.03088 fixed
sd_(Intercept).mother_pidlink 0.2591 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9609 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 7198 7262 -3588 7176 NA NA NA
12 7200 7270 -3588 7176 0.02429 1 0.8761
16 7200 7293 -3584 7168 8.315 4 0.08069
26 7214 7366 -3581 7162 5.651 10 0.8437

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -5.422 0.9015 -6.014 -7.188 -3.655 fixed
poly(age, 3, raw = TRUE)1 0.4572 0.09577 4.773 0.2694 0.6449 fixed
poly(age, 3, raw = TRUE)2 -0.01237 0.00326 -3.794 -0.01876 -0.005979 fixed
poly(age, 3, raw = TRUE)3 0.0001128 0.00003563 3.166 0.00004296 0.0001826 fixed
male 0.1739 0.04126 4.215 0.09305 0.2548 fixed
sibling_count3 -0.003189 0.06289 -0.05071 -0.1264 0.1201 fixed
sibling_count4 -0.1003 0.06552 -1.53 -0.2287 0.02815 fixed
sibling_count5 -0.1661 0.07676 -2.164 -0.3166 -0.01569 fixed
sibling_count5+ -0.1681 0.06519 -2.579 -0.2959 -0.04036 fixed
sd_(Intercept).mother_pidlink 0.2569 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.972 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -5.415 0.9017 -6.005 -7.182 -3.648 fixed
birth_order 0.008665 0.0132 0.6563 -0.01721 0.03455 fixed
poly(age, 3, raw = TRUE)1 0.4554 0.09582 4.753 0.2676 0.6432 fixed
poly(age, 3, raw = TRUE)2 -0.01233 0.003261 -3.78 -0.01872 -0.005935 fixed
poly(age, 3, raw = TRUE)3 0.0001127 0.00003563 3.161 0.00004281 0.0001825 fixed
male 0.1734 0.04128 4.2 0.09248 0.2543 fixed
sibling_count3 -0.007568 0.06326 -0.1196 -0.1315 0.1164 fixed
sibling_count4 -0.1098 0.06712 -1.635 -0.2413 0.02178 fixed
sibling_count5 -0.1818 0.08039 -2.261 -0.3393 -0.0242 fixed
sibling_count5+ -0.1994 0.08076 -2.469 -0.3577 -0.0411 fixed
sd_(Intercept).mother_pidlink 0.2582 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9718 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -5.41 0.9024 -5.996 -7.179 -3.642 fixed
poly(age, 3, raw = TRUE)1 0.4563 0.09589 4.759 0.2684 0.6443 fixed
poly(age, 3, raw = TRUE)2 -0.01236 0.003263 -3.788 -0.01876 -0.005966 fixed
poly(age, 3, raw = TRUE)3 0.0001131 0.00003566 3.171 0.00004319 0.000183 fixed
male 0.174 0.04128 4.214 0.09307 0.2549 fixed
sibling_count3 -0.02677 0.06466 -0.4139 -0.1535 0.09997 fixed
sibling_count4 -0.1287 0.06985 -1.842 -0.2656 0.008219 fixed
sibling_count5 -0.1758 0.08455 -2.079 -0.3415 -0.01005 fixed
sibling_count5+ -0.2091 0.08338 -2.508 -0.3726 -0.04571 fixed
birth_order_nonlinear2 -0.0007985 0.0533 -0.01498 -0.1053 0.1037 fixed
birth_order_nonlinear3 0.09794 0.06412 1.528 -0.02772 0.2236 fixed
birth_order_nonlinear4 0.0248 0.08029 0.3089 -0.1326 0.1822 fixed
birth_order_nonlinear5 -0.07846 0.1009 -0.7773 -0.2763 0.1194 fixed
birth_order_nonlinear5+ 0.1007 0.09905 1.016 -0.09349 0.2948 fixed
sd_(Intercept).mother_pidlink 0.2667 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9694 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -5.398 0.9042 -5.97 -7.171 -3.626 fixed
poly(age, 3, raw = TRUE)1 0.4538 0.09623 4.715 0.2652 0.6424 fixed
poly(age, 3, raw = TRUE)2 -0.01223 0.003276 -3.734 -0.01865 -0.005812 fixed
poly(age, 3, raw = TRUE)3 0.0001112 0.0000358 3.106 0.00004103 0.0001814 fixed
male 0.1673 0.04143 4.04 0.08615 0.2485 fixed
count_birth_order2/2 0.01444 0.09844 0.1467 -0.1785 0.2074 fixed
count_birth_order1/3 -0.06221 0.08283 -0.751 -0.2246 0.1001 fixed
count_birth_order2/3 0.07763 0.09437 0.8226 -0.1073 0.2626 fixed
count_birth_order3/3 0.03025 0.1018 0.297 -0.1694 0.2299 fixed
count_birth_order1/4 -0.1345 0.09873 -1.362 -0.328 0.05905 fixed
count_birth_order2/4 -0.1646 0.1011 -1.628 -0.3628 0.0336 fixed
count_birth_order3/4 0.02741 0.1071 0.2559 -0.1825 0.2373 fixed
count_birth_order4/4 -0.08879 0.1173 -0.757 -0.3187 0.1411 fixed
count_birth_order1/5 -0.002286 0.1351 -0.01692 -0.2671 0.2626 fixed
count_birth_order2/5 -0.1884 0.157 -1.2 -0.4961 0.1192 fixed
count_birth_order3/5 -0.1273 0.146 -0.8719 -0.4136 0.1589 fixed
count_birth_order4/5 -0.2074 0.1387 -1.496 -0.4792 0.0644 fixed
count_birth_order5/5 -0.3175 0.1436 -2.211 -0.5989 -0.03604 fixed
count_birth_order1/5+ -0.1604 0.1323 -1.212 -0.4196 0.09893 fixed
count_birth_order2/5+ -0.3924 0.1365 -2.875 -0.66 -0.1249 fixed
count_birth_order3/5+ -0.06706 0.134 -0.5004 -0.3297 0.1956 fixed
count_birth_order4/5+ -0.1438 0.1262 -1.14 -0.3911 0.1034 fixed
count_birth_order5/5+ -0.2396 0.1213 -1.974 -0.4774 -0.001747 fixed
count_birth_order5+/5+ -0.1037 0.09056 -1.145 -0.2812 0.0738 fixed
sd_(Intercept).mother_pidlink 0.261 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.9711 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 7066 7130 -3522 7044 NA NA NA
12 7067 7137 -3522 7043 0.4305 1 0.5117
16 7070 7163 -3519 7038 4.865 4 0.3014
26 7082 7233 -3515 7030 8.758 10 0.5552

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Self-Employment - non standardized

birthorder <- birthorder %>% mutate(outcome = Self_employed)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.2665 0.095 -2.805 -0.4527 -0.08029 fixed
poly(age, 3, raw = TRUE)1 0.02362 0.008521 2.772 0.006922 0.04033 fixed
poly(age, 3, raw = TRUE)2 -0.0001646 0.0002393 -0.6879 -0.0006336 0.0003044 fixed
poly(age, 3, raw = TRUE)3 -0.0000002153 0.000002126 -0.1013 -0.000004382 0.000003952 fixed
male -0.002977 0.008766 -0.3396 -0.02016 0.0142 fixed
sibling_count3 -0.02246 0.01828 -1.229 -0.05828 0.01337 fixed
sibling_count4 -0.01987 0.01844 -1.077 -0.056 0.01627 fixed
sibling_count5 -0.01023 0.01921 -0.5326 -0.04789 0.02743 fixed
sibling_count5+ 0.01732 0.0149 1.162 -0.01188 0.04651 fixed
sd_(Intercept).mother_pidlink 0.1531 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4098 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.2668 0.09499 -2.809 -0.453 -0.08063 fixed
birth_order 0.002029 0.001784 1.137 -0.001469 0.005526 fixed
poly(age, 3, raw = TRUE)1 0.02307 0.008535 2.703 0.006339 0.0398 fixed
poly(age, 3, raw = TRUE)2 -0.0001441 0.00024 -0.6006 -0.0006145 0.0003262 fixed
poly(age, 3, raw = TRUE)3 -0.0000004013 0.000002132 -0.1882 -0.00000458 0.000003778 fixed
male -0.003007 0.008766 -0.3431 -0.02019 0.01417 fixed
sibling_count3 -0.023 0.01828 -1.258 -0.05884 0.01283 fixed
sibling_count4 -0.0211 0.01847 -1.143 -0.0573 0.01509 fixed
sibling_count5 -0.01241 0.01931 -0.6426 -0.05025 0.02544 fixed
sibling_count5+ 0.01022 0.01615 0.6328 -0.02144 0.04188 fixed
sd_(Intercept).mother_pidlink 0.1529 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4098 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.2752 0.09511 -2.893 -0.4616 -0.08877 fixed
poly(age, 3, raw = TRUE)1 0.02319 0.008541 2.715 0.00645 0.03993 fixed
poly(age, 3, raw = TRUE)2 -0.0001598 0.0002401 -0.6657 -0.0006303 0.0003107 fixed
poly(age, 3, raw = TRUE)3 -0.0000001582 0.000002133 -0.07415 -0.000004338 0.000004022 fixed
male -0.003111 0.008765 -0.3549 -0.02029 0.01407 fixed
sibling_count3 -0.02462 0.01855 -1.327 -0.06098 0.01173 fixed
sibling_count4 -0.02199 0.01896 -1.16 -0.05915 0.01517 fixed
sibling_count5 -0.01296 0.01996 -0.6493 -0.05208 0.02616 fixed
sibling_count5+ 0.01309 0.01691 0.7741 -0.02005 0.04623 fixed
birth_order_nonlinear2 0.03823 0.01276 2.997 0.01323 0.06324 fixed
birth_order_nonlinear3 0.01961 0.01483 1.322 -0.009457 0.04867 fixed
birth_order_nonlinear4 0.0146 0.01665 0.877 -0.01803 0.04724 fixed
birth_order_nonlinear5 0.01968 0.01874 1.05 -0.01704 0.0564 fixed
birth_order_nonlinear5+ 0.02218 0.01558 1.423 -0.008359 0.05271 fixed
sd_(Intercept).mother_pidlink 0.1529 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4097 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.2834 0.09546 -2.969 -0.4705 -0.09631 fixed
poly(age, 3, raw = TRUE)1 0.02291 0.008548 2.68 0.006152 0.03966 fixed
poly(age, 3, raw = TRUE)2 -0.0001582 0.0002402 -0.6585 -0.000629 0.0003126 fixed
poly(age, 3, raw = TRUE)3 -0.000000111 0.000002134 -0.052 -0.000004293 0.000004072 fixed
male -0.003372 0.008769 -0.3845 -0.02056 0.01382 fixed
count_birth_order2/2 0.07421 0.02534 2.928 0.02454 0.1239 fixed
count_birth_order1/3 -0.02418 0.02456 -0.9844 -0.07231 0.02396 fixed
count_birth_order2/3 0.03323 0.02722 1.221 -0.02011 0.08657 fixed
count_birth_order3/3 0.02459 0.02989 0.8229 -0.03398 0.08317 fixed
count_birth_order1/4 0.01501 0.0269 0.5579 -0.03772 0.06773 fixed
count_birth_order2/4 0.005201 0.02884 0.1804 -0.05132 0.06173 fixed
count_birth_order3/4 -0.005191 0.03075 -0.1688 -0.06545 0.05507 fixed
count_birth_order4/4 0.01914 0.03266 0.5862 -0.04486 0.08315 fixed
count_birth_order1/5 0.009117 0.03054 0.2985 -0.05074 0.06897 fixed
count_birth_order2/5 0.04643 0.03238 1.434 -0.01704 0.1099 fixed
count_birth_order3/5 0.03461 0.03409 1.015 -0.0322 0.1014 fixed
count_birth_order4/5 -0.01429 0.03583 -0.3989 -0.08451 0.05593 fixed
count_birth_order5/5 0.0102 0.03582 0.2846 -0.06001 0.0804 fixed
count_birth_order1/5+ 0.03492 0.02352 1.485 -0.01118 0.08102 fixed
count_birth_order2/5+ 0.05428 0.02449 2.216 0.006268 0.1023 fixed
count_birth_order3/5+ 0.04115 0.02413 1.705 -0.006149 0.08844 fixed
count_birth_order4/5+ 0.04521 0.0237 1.907 -0.001244 0.09166 fixed
count_birth_order5/5+ 0.04958 0.02396 2.069 0.002613 0.09655 fixed
count_birth_order5+/5+ 0.04898 0.01941 2.523 0.01094 0.08703 fixed
sd_(Intercept).mother_pidlink 0.153 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4097 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 11871 11950 -5924 11849 NA NA NA
12 11871 11958 -5924 11847 1.295 1 0.2552
16 11872 11987 -5920 11840 7.919 4 0.0946
26 11883 12071 -5916 11831 8.217 10 0.6077

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.9428 0.2612 3.609 0.4307 1.455 fixed
poly(age, 3, raw = TRUE)1 -0.1058 0.02805 -3.773 -0.1608 -0.05087 fixed
poly(age, 3, raw = TRUE)2 0.004147 0.0009623 4.309 0.002261 0.006033 fixed
poly(age, 3, raw = TRUE)3 -0.00004614 0.00001056 -4.368 -0.00006685 -0.00002544 fixed
male -0.02367 0.01315 -1.8 -0.04943 0.0021 fixed
sibling_count3 0.01552 0.02194 0.7075 -0.02747 0.05851 fixed
sibling_count4 -0.02907 0.02271 -1.28 -0.07358 0.01545 fixed
sibling_count5 0.03856 0.02531 1.524 -0.01104 0.08815 fixed
sibling_count5+ 0.04773 0.02198 2.172 0.00465 0.09081 fixed
sd_(Intercept).mother_pidlink 0.1578 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3721 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.9413 0.2613 3.602 0.4291 1.453 fixed
birth_order -0.001447 0.004107 -0.3524 -0.009497 0.006603 fixed
poly(age, 3, raw = TRUE)1 -0.1055 0.02807 -3.759 -0.1605 -0.05049 fixed
poly(age, 3, raw = TRUE)2 0.004138 0.0009626 4.299 0.002251 0.006025 fixed
poly(age, 3, raw = TRUE)3 -0.00004611 0.00001057 -4.364 -0.00006682 -0.0000254 fixed
male -0.02362 0.01315 -1.796 -0.04939 0.002152 fixed
sibling_count3 0.01627 0.02204 0.7382 -0.02693 0.05948 fixed
sibling_count4 -0.02744 0.02318 -1.184 -0.07288 0.01799 fixed
sibling_count5 0.0412 0.02642 1.559 -0.01058 0.09299 fixed
sibling_count5+ 0.05295 0.02654 1.995 0.0009406 0.105 fixed
sd_(Intercept).mother_pidlink 0.1582 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.372 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.9329 0.2617 3.565 0.42 1.446 fixed
poly(age, 3, raw = TRUE)1 -0.1051 0.02809 -3.742 -0.1602 -0.05006 fixed
poly(age, 3, raw = TRUE)2 0.004127 0.0009634 4.283 0.002238 0.006015 fixed
poly(age, 3, raw = TRUE)3 -0.00004607 0.00001058 -4.355 -0.0000668 -0.00002534 fixed
male -0.02352 0.01315 -1.788 -0.0493 0.002267 fixed
sibling_count3 0.0156 0.02247 0.6942 -0.02844 0.05964 fixed
sibling_count4 -0.02767 0.02398 -1.154 -0.07467 0.01933 fixed
sibling_count5 0.04284 0.02763 1.55 -0.01132 0.09699 fixed
sibling_count5+ 0.0613 0.02722 2.252 0.007956 0.1147 fixed
birth_order_nonlinear2 0.01025 0.01727 0.5934 -0.02361 0.04411 fixed
birth_order_nonlinear3 0.001374 0.02033 0.06759 -0.03848 0.04123 fixed
birth_order_nonlinear4 -0.002639 0.02479 -0.1065 -0.05122 0.04594 fixed
birth_order_nonlinear5 -0.008891 0.03048 -0.2917 -0.06864 0.05086 fixed
birth_order_nonlinear5+ -0.02557 0.03043 -0.8404 -0.08522 0.03407 fixed
sd_(Intercept).mother_pidlink 0.1581 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3722 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.9316 0.2628 3.545 0.4166 1.447 fixed
poly(age, 3, raw = TRUE)1 -0.1065 0.02821 -3.774 -0.1618 -0.05117 fixed
poly(age, 3, raw = TRUE)2 0.004174 0.0009678 4.313 0.002277 0.006071 fixed
poly(age, 3, raw = TRUE)3 -0.00004658 0.00001063 -4.383 -0.00006741 -0.00002575 fixed
male -0.02415 0.01318 -1.832 -0.04998 0.001684 fixed
count_birth_order2/2 0.05231 0.0336 1.557 -0.01355 0.1182 fixed
count_birth_order1/3 0.03981 0.02861 1.392 -0.01626 0.09589 fixed
count_birth_order2/3 0.03881 0.03176 1.222 -0.02344 0.1011 fixed
count_birth_order3/3 0.01045 0.03434 0.3044 -0.05685 0.07775 fixed
count_birth_order1/4 -0.008259 0.03285 -0.2514 -0.07264 0.05612 fixed
count_birth_order2/4 -0.02985 0.03453 -0.8646 -0.09752 0.03782 fixed
count_birth_order3/4 -0.01272 0.0362 -0.3513 -0.08366 0.05823 fixed
count_birth_order4/4 0.008705 0.03839 0.2267 -0.06654 0.08395 fixed
count_birth_order1/5 0.05321 0.04307 1.235 -0.03121 0.1376 fixed
count_birth_order2/5 0.07654 0.04684 1.634 -0.01527 0.1684 fixed
count_birth_order3/5 0.06749 0.04389 1.538 -0.01854 0.1535 fixed
count_birth_order4/5 0.05473 0.04269 1.282 -0.02895 0.1384 fixed
count_birth_order5/5 0.02978 0.04476 0.6652 -0.05796 0.1175 fixed
count_birth_order1/5+ 0.07785 0.04095 1.901 -0.002414 0.1581 fixed
count_birth_order2/5+ 0.07465 0.04131 1.807 -0.006317 0.1556 fixed
count_birth_order3/5+ 0.09799 0.04038 2.426 0.01884 0.1771 fixed
count_birth_order4/5+ 0.04506 0.03922 1.149 -0.0318 0.1219 fixed
count_birth_order5/5+ 0.07709 0.03782 2.038 0.00296 0.1512 fixed
count_birth_order5+/5+ 0.04905 0.0296 1.657 -0.008953 0.1071 fixed
sd_(Intercept).mother_pidlink 0.1579 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3725 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 3908 3977 -1943 3886 NA NA NA
12 3910 3985 -1943 3886 0.122 1 0.7269
16 3917 4017 -1942 3885 1.352 4 0.8525
26 3931 4094 -1940 3879 5.638 10 0.8447

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.917 0.2603 3.523 0.4068 1.427 fixed
poly(age, 3, raw = TRUE)1 -0.1019 0.02798 -3.642 -0.1567 -0.04707 fixed
poly(age, 3, raw = TRUE)2 0.004027 0.0009598 4.196 0.002146 0.005909 fixed
poly(age, 3, raw = TRUE)3 -0.00004496 0.00001054 -4.266 -0.00006562 -0.0000243 fixed
male -0.02421 0.01311 -1.847 -0.04991 0.001481 fixed
sibling_count3 -0.006695 0.02405 -0.2784 -0.05383 0.04044 fixed
sibling_count4 -0.03614 0.02428 -1.489 -0.08372 0.01144 fixed
sibling_count5 -0.008562 0.02563 -0.3341 -0.05879 0.04167 fixed
sibling_count5+ 0.0278 0.02245 1.238 -0.0162 0.0718 fixed
sd_(Intercept).mother_pidlink 0.1587 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.372 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.9174 0.2603 3.524 0.4071 1.428 fixed
birth_order 0.0004435 0.003628 0.1222 -0.006668 0.007555 fixed
poly(age, 3, raw = TRUE)1 -0.102 0.02799 -3.644 -0.1568 -0.04713 fixed
poly(age, 3, raw = TRUE)2 0.00403 0.0009601 4.197 0.002148 0.005911 fixed
poly(age, 3, raw = TRUE)3 -0.00004497 0.00001054 -4.266 -0.00006563 -0.00002431 fixed
male -0.02422 0.01311 -1.848 -0.04992 0.001474 fixed
sibling_count3 -0.006931 0.02413 -0.2872 -0.05422 0.04036 fixed
sibling_count4 -0.03661 0.02457 -1.49 -0.08477 0.01156 fixed
sibling_count5 -0.009323 0.02638 -0.3534 -0.06102 0.04238 fixed
sibling_count5+ 0.02623 0.02589 1.013 -0.02451 0.07696 fixed
sd_(Intercept).mother_pidlink 0.1587 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3721 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.9044 0.2607 3.47 0.3935 1.415 fixed
poly(age, 3, raw = TRUE)1 -0.1011 0.02801 -3.608 -0.156 -0.04616 fixed
poly(age, 3, raw = TRUE)2 0.003996 0.0009609 4.159 0.002113 0.00588 fixed
poly(age, 3, raw = TRUE)3 -0.00004462 0.00001055 -4.229 -0.00006531 -0.00002394 fixed
male -0.02403 0.01312 -1.832 -0.04974 0.001677 fixed
sibling_count3 -0.01015 0.02454 -0.4134 -0.05824 0.03795 fixed
sibling_count4 -0.03844 0.02528 -1.521 -0.08799 0.0111 fixed
sibling_count5 -0.008874 0.02749 -0.3228 -0.06275 0.04501 fixed
sibling_count5+ 0.02904 0.02661 1.091 -0.02311 0.08119 fixed
birth_order_nonlinear2 0.01753 0.01751 1.001 -0.01679 0.05185 fixed
birth_order_nonlinear3 0.01558 0.02045 0.7616 -0.02451 0.05566 fixed
birth_order_nonlinear4 -0.00005852 0.0242 -0.002418 -0.04749 0.04737 fixed
birth_order_nonlinear5 -0.002062 0.02949 -0.06994 -0.05986 0.05573 fixed
birth_order_nonlinear5+ 0.002764 0.02752 0.1004 -0.05118 0.05671 fixed
sd_(Intercept).mother_pidlink 0.1585 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3722 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.904 0.2612 3.46 0.392 1.416 fixed
poly(age, 3, raw = TRUE)1 -0.1033 0.02807 -3.68 -0.1583 -0.04828 fixed
poly(age, 3, raw = TRUE)2 0.004077 0.0009634 4.233 0.002189 0.005966 fixed
poly(age, 3, raw = TRUE)3 -0.00004555 0.00001058 -4.305 -0.00006629 -0.00002481 fixed
male -0.02482 0.01313 -1.89 -0.05057 0.0009199 fixed
count_birth_order2/2 0.07947 0.03699 2.148 0.006965 0.152 fixed
count_birth_order1/3 0.02032 0.0316 0.643 -0.04161 0.08225 fixed
count_birth_order2/3 0.04303 0.03437 1.252 -0.02433 0.1104 fixed
count_birth_order3/3 -0.015 0.03778 -0.3971 -0.08904 0.05904 fixed
count_birth_order1/4 0.005795 0.03435 0.1687 -0.06153 0.07312 fixed
count_birth_order2/4 -0.0307 0.03584 -0.8566 -0.1009 0.03954 fixed
count_birth_order3/4 -0.0108 0.03917 -0.2758 -0.08757 0.06596 fixed
count_birth_order4/4 -0.006404 0.04181 -0.1532 -0.08835 0.07554 fixed
count_birth_order1/5 0.01609 0.04117 0.3908 -0.0646 0.09678 fixed
count_birth_order2/5 0.01692 0.04276 0.3957 -0.06688 0.1007 fixed
count_birth_order3/5 0.01622 0.04352 0.3728 -0.06907 0.1015 fixed
count_birth_order4/5 0.01776 0.04424 0.4013 -0.06896 0.1045 fixed
count_birth_order5/5 0.02105 0.04457 0.4723 -0.0663 0.1084 fixed
count_birth_order1/5+ 0.03557 0.03723 0.9554 -0.0374 0.1085 fixed
count_birth_order2/5+ 0.04765 0.03933 1.211 -0.02944 0.1247 fixed
count_birth_order3/5+ 0.1171 0.03789 3.092 0.04288 0.1914 fixed
count_birth_order4/5+ 0.03739 0.03695 1.012 -0.03504 0.1098 fixed
count_birth_order5/5+ 0.03913 0.03908 1.001 -0.03746 0.1157 fixed
count_birth_order5+/5+ 0.05174 0.02961 1.747 -0.006299 0.1098 fixed
sd_(Intercept).mother_pidlink 0.1582 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3723 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 3939 4008 -1958 3917 NA NA NA
12 3941 4016 -1958 3917 0.01549 1 0.901
16 3947 4047 -1958 3915 1.543 4 0.8189
26 3955 4118 -1952 3903 11.8 10 0.2988

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.9712 0.2646 3.67 0.4525 1.49 fixed
poly(age, 3, raw = TRUE)1 -0.11 0.02846 -3.864 -0.1657 -0.05418 fixed
poly(age, 3, raw = TRUE)2 0.004334 0.0009771 4.435 0.002418 0.006249 fixed
poly(age, 3, raw = TRUE)3 -0.00004871 0.00001074 -4.536 -0.00006976 -0.00002766 fixed
male -0.02276 0.0133 -1.711 -0.04883 0.003314 fixed
sibling_count3 0.02252 0.02157 1.044 -0.01976 0.06479 fixed
sibling_count4 -0.02657 0.02265 -1.173 -0.07097 0.01783 fixed
sibling_count5 0.03133 0.02594 1.208 -0.01952 0.08218 fixed
sibling_count5+ 0.04105 0.02217 1.852 -0.002396 0.0845 fixed
sd_(Intercept).mother_pidlink 0.16 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3726 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.9698 0.2647 3.663 0.4509 1.489 fixed
birth_order -0.001067 0.004238 -0.2517 -0.009373 0.007239 fixed
poly(age, 3, raw = TRUE)1 -0.1097 0.02848 -3.851 -0.1655 -0.05385 fixed
poly(age, 3, raw = TRUE)2 0.004326 0.0009776 4.425 0.00241 0.006242 fixed
poly(age, 3, raw = TRUE)3 -0.00004867 0.00001074 -4.531 -0.00006973 -0.00002762 fixed
male -0.02275 0.0133 -1.71 -0.04882 0.00333 fixed
sibling_count3 0.02308 0.02169 1.064 -0.01943 0.0656 fixed
sibling_count4 -0.02539 0.02314 -1.097 -0.07074 0.01996 fixed
sibling_count5 0.03322 0.02704 1.229 -0.01977 0.08622 fixed
sibling_count5+ 0.0449 0.02695 1.666 -0.007923 0.09772 fixed
sd_(Intercept).mother_pidlink 0.1602 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3725 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.9616 0.2651 3.628 0.4421 1.481 fixed
poly(age, 3, raw = TRUE)1 -0.1091 0.0285 -3.828 -0.165 -0.05323 fixed
poly(age, 3, raw = TRUE)2 0.00431 0.0009784 4.405 0.002392 0.006227 fixed
poly(age, 3, raw = TRUE)3 -0.00004859 0.00001075 -4.519 -0.00006966 -0.00002751 fixed
male -0.02281 0.01331 -1.713 -0.0489 0.003284 fixed
sibling_count3 0.02313 0.02213 1.045 -0.02025 0.06651 fixed
sibling_count4 -0.02585 0.02394 -1.08 -0.07278 0.02107 fixed
sibling_count5 0.03196 0.02818 1.134 -0.02328 0.08719 fixed
sibling_count5+ 0.05248 0.02768 1.896 -0.001773 0.1067 fixed
birth_order_nonlinear2 0.004606 0.01724 0.2672 -0.02918 0.03839 fixed
birth_order_nonlinear3 -0.001029 0.02041 -0.0504 -0.04103 0.03897 fixed
birth_order_nonlinear4 0.001722 0.02559 0.06727 -0.04844 0.05189 fixed
birth_order_nonlinear5 0.003465 0.03152 0.1099 -0.05832 0.06525 fixed
birth_order_nonlinear5+ -0.02943 0.03147 -0.9352 -0.09112 0.03225 fixed
sd_(Intercept).mother_pidlink 0.1602 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3727 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.9503 0.266 3.573 0.429 1.472 fixed
poly(age, 3, raw = TRUE)1 -0.1092 0.0286 -3.819 -0.1653 -0.05318 fixed
poly(age, 3, raw = TRUE)2 0.004316 0.0009823 4.394 0.002391 0.006241 fixed
poly(age, 3, raw = TRUE)3 -0.00004867 0.0000108 -4.507 -0.00006983 -0.00002751 fixed
male -0.0234 0.01334 -1.754 -0.04954 0.002743 fixed
count_birth_order2/2 0.0426 0.03271 1.302 -0.02152 0.1067 fixed
count_birth_order1/3 0.04535 0.0282 1.608 -0.009916 0.1006 fixed
count_birth_order2/3 0.04406 0.03134 1.406 -0.01736 0.1055 fixed
count_birth_order3/3 0.01164 0.03353 0.3472 -0.05408 0.07737 fixed
count_birth_order1/4 -0.008358 0.03305 -0.2528 -0.07314 0.05643 fixed
count_birth_order2/4 -0.04096 0.03452 -1.187 -0.1086 0.0267 fixed
count_birth_order3/4 -0.01007 0.03618 -0.2784 -0.08097 0.06083 fixed
count_birth_order4/4 0.02124 0.03901 0.5443 -0.05522 0.09769 fixed
count_birth_order1/5 0.0574 0.04322 1.328 -0.02731 0.1421 fixed
count_birth_order2/5 0.07243 0.04845 1.495 -0.02254 0.1674 fixed
count_birth_order3/5 0.0487 0.04642 1.049 -0.04227 0.1397 fixed
count_birth_order4/5 0.03821 0.04477 0.8534 -0.04954 0.126 fixed
count_birth_order5/5 0.0122 0.04712 0.2589 -0.08016 0.1046 fixed
count_birth_order1/5+ 0.05652 0.0421 1.342 -0.026 0.139 fixed
count_birth_order2/5+ 0.05222 0.04261 1.225 -0.0313 0.1357 fixed
count_birth_order3/5+ 0.09224 0.04079 2.261 0.01228 0.1722 fixed
count_birth_order4/5+ 0.03775 0.04057 0.9304 -0.04177 0.1173 fixed
count_birth_order5/5+ 0.0893 0.0384 2.326 0.01404 0.1646 fixed
count_birth_order5+/5+ 0.0354 0.0301 1.176 -0.02359 0.09438 fixed
sd_(Intercept).mother_pidlink 0.1598 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3729 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 3861 3930 -1919 3839 NA NA NA
12 3863 3938 -1919 3839 0.06169 1 0.8038
16 3870 3969 -1919 3838 1.332 4 0.856
26 3882 4044 -1915 3830 8.08 10 0.6211

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Work Category

Category_Casual worker in agriculture

birthorder <- birthorder %>% mutate(outcome = `Category_Casual worker in agriculture`)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.06852 0.03525 1.944 -0.0005704 0.1376 fixed
poly(age, 3, raw = TRUE)1 -0.004585 0.003161 -1.45 -0.01078 0.001611 fixed
poly(age, 3, raw = TRUE)2 0.0001275 0.00008876 1.437 -0.00004643 0.0003015 fixed
poly(age, 3, raw = TRUE)3 -0.0000009703 0.0000007884 -1.231 -0.000002516 0.000000575 fixed
male 0.0003999 0.003258 0.1228 -0.005985 0.006785 fixed
sibling_count3 -0.00296 0.006763 -0.4378 -0.01622 0.01029 fixed
sibling_count4 0.005901 0.00682 0.8653 -0.007465 0.01927 fixed
sibling_count5 -0.002795 0.007104 -0.3935 -0.01672 0.01113 fixed
sibling_count5+ 0.008566 0.005512 1.554 -0.002237 0.01937 fixed
sd_(Intercept).mother_pidlink 0.05392 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1531 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.06862 0.03525 1.947 -0.0004662 0.1377 fixed
birth_order -0.0009169 0.0006614 -1.386 -0.002213 0.0003793 fixed
poly(age, 3, raw = TRUE)1 -0.004333 0.003167 -1.368 -0.01054 0.001873 fixed
poly(age, 3, raw = TRUE)2 0.0001183 0.00008901 1.329 -0.00005613 0.0002928 fixed
poly(age, 3, raw = TRUE)3 -0.0000008869 0.0000007907 -1.122 -0.000002437 0.0000006629 fixed
male 0.0004145 0.003257 0.1273 -0.00597 0.006799 fixed
sibling_count3 -0.00271 0.006766 -0.4005 -0.01597 0.01055 fixed
sibling_count4 0.006467 0.006833 0.9464 -0.006925 0.01986 fixed
sibling_count5 -0.001804 0.007141 -0.2526 -0.0158 0.01219 fixed
sibling_count5+ 0.01179 0.005981 1.971 0.00006478 0.02351 fixed
sd_(Intercept).mother_pidlink 0.05405 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.153 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.0684 0.03531 1.937 -0.000797 0.1376 fixed
poly(age, 3, raw = TRUE)1 -0.004322 0.00317 -1.364 -0.01053 0.001891 fixed
poly(age, 3, raw = TRUE)2 0.0001208 0.00008907 1.357 -0.00005374 0.0002954 fixed
poly(age, 3, raw = TRUE)3 -0.0000009306 0.0000007912 -1.176 -0.000002481 0.0000006201 fixed
male 0.0004062 0.003258 0.1247 -0.00598 0.006792 fixed
sibling_count3 -0.002613 0.006868 -0.3805 -0.01607 0.01085 fixed
sibling_count4 0.007246 0.007019 1.032 -0.006511 0.021 fixed
sibling_count5 -0.001002 0.007388 -0.1356 -0.01548 0.01348 fixed
sibling_count5+ 0.01129 0.006267 1.801 -0.0009978 0.02357 fixed
birth_order_nonlinear2 -0.00616 0.004749 -1.297 -0.01547 0.003147 fixed
birth_order_nonlinear3 -0.003442 0.00552 -0.6236 -0.01426 0.007377 fixed
birth_order_nonlinear4 -0.008114 0.006198 -1.309 -0.02026 0.004034 fixed
birth_order_nonlinear5 -0.006216 0.006973 -0.8915 -0.01988 0.007451 fixed
birth_order_nonlinear5+ -0.006247 0.005788 -1.079 -0.01759 0.005097 fixed
sd_(Intercept).mother_pidlink 0.05398 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1531 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.07183 0.03542 2.028 0.002406 0.1413 fixed
poly(age, 3, raw = TRUE)1 -0.004317 0.003172 -1.361 -0.01053 0.001899 fixed
poly(age, 3, raw = TRUE)2 0.0001217 0.0000891 1.366 -0.00005296 0.0002963 fixed
poly(age, 3, raw = TRUE)3 -0.0000009483 0.0000007913 -1.198 -0.000002499 0.0000006027 fixed
male 0.0003518 0.003259 0.1079 -0.006036 0.00674 fixed
count_birth_order2/2 -0.01611 0.009432 -1.708 -0.03459 0.002377 fixed
count_birth_order1/3 -0.007027 0.009116 -0.7708 -0.02489 0.01084 fixed
count_birth_order2/3 -0.006682 0.0101 -0.6613 -0.02649 0.01312 fixed
count_birth_order3/3 -0.01617 0.0111 -1.457 -0.03792 0.005584 fixed
count_birth_order1/4 -0.007734 0.009987 -0.7744 -0.02731 0.01184 fixed
count_birth_order2/4 -0.003666 0.01071 -0.3423 -0.02465 0.01732 fixed
count_birth_order3/4 0.01068 0.01142 0.935 -0.0117 0.03305 fixed
count_birth_order4/4 0.00423 0.01213 0.3488 -0.01954 0.028 fixed
count_birth_order1/5 0.002099 0.01134 0.1851 -0.02013 0.02433 fixed
count_birth_order2/5 -0.01472 0.01203 -1.224 -0.03829 0.008845 fixed
count_birth_order3/5 -0.02292 0.01266 -1.81 -0.04773 0.001896 fixed
count_birth_order4/5 0.003402 0.01331 0.2556 -0.02268 0.02948 fixed
count_birth_order5/5 -0.01578 0.0133 -1.186 -0.04186 0.01029 fixed
count_birth_order1/5+ 0.007944 0.008735 0.9094 -0.009177 0.02506 fixed
count_birth_order2/5+ 0.004563 0.009098 0.5016 -0.01327 0.02239 fixed
count_birth_order3/5+ 0.007596 0.008963 0.8475 -0.009971 0.02516 fixed
count_birth_order4/5+ -0.008504 0.008802 -0.9661 -0.02576 0.008748 fixed
count_birth_order5/5+ 0.002768 0.008901 0.311 -0.01468 0.02021 fixed
count_birth_order5+/5+ 0.001264 0.007196 0.1757 -0.01284 0.01537 fixed
sd_(Intercept).mother_pidlink 0.05386 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1531 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -8128 -8049 4075 -8150 NA NA NA
12 -8128 -8041 4076 -8152 1.92 1 0.1659
16 -8121 -8005 4076 -8153 0.7708 4 0.9423
26 -8114 -7926 4083 -8166 13.04 10 0.2216

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.01952 0.07867 0.2481 -0.1347 0.1737 fixed
poly(age, 3, raw = TRUE)1 0.001255 0.008445 0.1486 -0.0153 0.01781 fixed
poly(age, 3, raw = TRUE)2 -0.0001097 0.0002896 -0.3787 -0.0006773 0.000458 fixed
poly(age, 3, raw = TRUE)3 0.000001805 0.000003179 0.5676 -0.000004427 0.000008036 fixed
male 0.003915 0.003952 0.9909 -0.003829 0.01166 fixed
sibling_count3 -0.005883 0.006687 -0.8798 -0.01899 0.007224 fixed
sibling_count4 0.0002679 0.006935 0.03863 -0.01332 0.01386 fixed
sibling_count5 -0.007085 0.007746 -0.9147 -0.02227 0.008096 fixed
sibling_count5+ 0.01689 0.006728 2.51 0.003703 0.03008 fixed
sd_(Intercept).mother_pidlink 0.05521 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1091 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.02063 0.07869 0.2622 -0.1336 0.1749 fixed
birth_order 0.0009961 0.001241 0.803 -0.001435 0.003427 fixed
poly(age, 3, raw = TRUE)1 0.001006 0.008451 0.119 -0.01556 0.01757 fixed
poly(age, 3, raw = TRUE)2 -0.0001032 0.0002898 -0.3563 -0.0006711 0.0004647 fixed
poly(age, 3, raw = TRUE)3 0.000001779 0.00000318 0.5595 -0.000004453 0.000008011 fixed
male 0.003886 0.003952 0.9833 -0.00386 0.01163 fixed
sibling_count3 -0.006403 0.006718 -0.953 -0.01957 0.006765 fixed
sibling_count4 -0.0008584 0.007075 -0.1213 -0.01473 0.01301 fixed
sibling_count5 -0.008934 0.00808 -1.106 -0.02477 0.006903 fixed
sibling_count5+ 0.01326 0.008107 1.635 -0.002634 0.02915 fixed
sd_(Intercept).mother_pidlink 0.05515 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1091 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.02295 0.07878 0.2913 -0.1315 0.1773 fixed
poly(age, 3, raw = TRUE)1 0.0008336 0.008454 0.0986 -0.01574 0.0174 fixed
poly(age, 3, raw = TRUE)2 -0.0000993 0.0002899 -0.3425 -0.0006675 0.0004689 fixed
poly(age, 3, raw = TRUE)3 0.000001774 0.000003182 0.5574 -0.000004463 0.00000801 fixed
male 0.003896 0.003953 0.9856 -0.003851 0.01164 fixed
sibling_count3 -0.005982 0.006841 -0.8744 -0.01939 0.007427 fixed
sibling_count4 -0.0002267 0.007308 -0.03102 -0.01455 0.0141 fixed
sibling_count5 -0.00734 0.008432 -0.8706 -0.02387 0.009185 fixed
sibling_count5+ 0.01131 0.008306 1.361 -0.004973 0.02758 fixed
birth_order_nonlinear2 0.002468 0.005159 0.4784 -0.007644 0.01258 fixed
birth_order_nonlinear3 0.0001147 0.006077 0.01887 -0.0118 0.01203 fixed
birth_order_nonlinear4 0.001705 0.007413 0.23 -0.01282 0.01623 fixed
birth_order_nonlinear5 -0.0005972 0.009118 -0.06549 -0.01847 0.01727 fixed
birth_order_nonlinear5+ 0.0157 0.009163 1.713 -0.002263 0.03365 fixed
sd_(Intercept).mother_pidlink 0.05527 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1091 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.02181 0.07903 0.276 -0.1331 0.1767 fixed
poly(age, 3, raw = TRUE)1 0.001327 0.008482 0.1565 -0.0153 0.01795 fixed
poly(age, 3, raw = TRUE)2 -0.0001173 0.0002909 -0.4032 -0.0006875 0.0004529 fixed
poly(age, 3, raw = TRUE)3 0.00000199 0.000003194 0.6231 -0.00000427 0.000008251 fixed
male 0.003913 0.003957 0.989 -0.003842 0.01167 fixed
count_birth_order2/2 -0.007885 0.01004 -0.785 -0.02757 0.0118 fixed
count_birth_order1/3 -0.01161 0.008624 -1.346 -0.02851 0.005296 fixed
count_birth_order2/3 -0.009642 0.009564 -1.008 -0.02839 0.009103 fixed
count_birth_order3/3 -0.001103 0.01033 -0.1067 -0.02136 0.01915 fixed
count_birth_order1/4 0.001448 0.009895 0.1463 -0.01795 0.02084 fixed
count_birth_order2/4 -0.001238 0.01039 -0.1191 -0.02161 0.01913 fixed
count_birth_order3/4 -0.0139 0.01089 -1.277 -0.03524 0.007441 fixed
count_birth_order4/4 0.002229 0.01154 0.1931 -0.0204 0.02486 fixed
count_birth_order1/5 -0.009352 0.01296 -0.7215 -0.03475 0.01605 fixed
count_birth_order2/5 -0.01508 0.01408 -1.071 -0.04266 0.01251 fixed
count_birth_order3/5 -0.01413 0.01319 -1.071 -0.03998 0.01173 fixed
count_birth_order4/5 0.002559 0.01283 0.1994 -0.02259 0.02771 fixed
count_birth_order5/5 -0.01591 0.01345 -1.183 -0.04227 0.01046 fixed
count_birth_order1/5+ -0.00645 0.01232 -0.5237 -0.03059 0.01769 fixed
count_birth_order2/5+ 0.03205 0.01241 2.582 0.007721 0.05638 fixed
count_birth_order3/5+ 0.01239 0.01213 1.022 -0.01138 0.03617 fixed
count_birth_order4/5+ -0.003848 0.01178 -0.3267 -0.02693 0.01924 fixed
count_birth_order5/5+ 0.01032 0.01136 0.9085 -0.01195 0.03259 fixed
count_birth_order5+/5+ 0.02363 0.008944 2.642 0.006101 0.04116 fixed
sd_(Intercept).mother_pidlink 0.05466 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1093 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -5280 -5211 2651 -5302 NA NA NA
12 -5279 -5203 2651 -5303 0.6479 1 0.4209
16 -5274 -5174 2653 -5306 3.207 4 0.5238
26 -5268 -5106 2660 -5320 14.78 10 0.1402

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.01131 0.0782 0.1446 -0.142 0.1646 fixed
poly(age, 3, raw = TRUE)1 0.002056 0.008402 0.2447 -0.01441 0.01852 fixed
poly(age, 3, raw = TRUE)2 -0.0001338 0.0002882 -0.4642 -0.0006987 0.0004311 fixed
poly(age, 3, raw = TRUE)3 0.000002068 0.000003165 0.6536 -0.000004134 0.000008271 fixed
male 0.003961 0.003931 1.007 -0.003744 0.01167 fixed
sibling_count3 -0.003733 0.007306 -0.5109 -0.01805 0.01059 fixed
sibling_count4 -0.002881 0.007382 -0.3903 -0.01735 0.01159 fixed
sibling_count5 -0.005224 0.007806 -0.6693 -0.02052 0.01007 fixed
sibling_count5+ 0.00942 0.006833 1.379 -0.003973 0.02281 fixed
sd_(Intercept).mother_pidlink 0.05499 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1089 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.01302 0.07819 0.1665 -0.1402 0.1663 fixed
birth_order 0.001766 0.001094 1.613 -0.0003795 0.003911 fixed
poly(age, 3, raw = TRUE)1 0.001647 0.008404 0.1959 -0.01483 0.01812 fixed
poly(age, 3, raw = TRUE)2 -0.0001235 0.0002882 -0.4285 -0.0006885 0.0004414 fixed
poly(age, 3, raw = TRUE)3 0.000002037 0.000003164 0.6439 -0.000004164 0.000008239 fixed
male 0.003919 0.003931 0.997 -0.003785 0.01162 fixed
sibling_count3 -0.004682 0.007326 -0.6391 -0.01904 0.009677 fixed
sibling_count4 -0.004751 0.007469 -0.6361 -0.01939 0.009888 fixed
sibling_count5 -0.00828 0.008028 -1.031 -0.02402 0.007456 fixed
sibling_count5+ 0.00311 0.00787 0.3951 -0.01232 0.01853 fixed
sd_(Intercept).mother_pidlink 0.0548 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.109 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.01294 0.07828 0.1654 -0.1405 0.1664 fixed
poly(age, 3, raw = TRUE)1 0.001859 0.008409 0.2211 -0.01462 0.01834 fixed
poly(age, 3, raw = TRUE)2 -0.0001321 0.0002884 -0.4581 -0.0006975 0.0004332 fixed
poly(age, 3, raw = TRUE)3 0.00000215 0.000003167 0.6788 -0.000004058 0.000008357 fixed
male 0.003953 0.003932 1.005 -0.003754 0.01166 fixed
sibling_count3 -0.003594 0.007444 -0.4828 -0.01818 0.011 fixed
sibling_count4 -0.002937 0.007673 -0.3828 -0.01798 0.0121 fixed
sibling_count5 -0.006432 0.008351 -0.7703 -0.0228 0.009935 fixed
sibling_count5+ 0.00298 0.008078 0.3689 -0.01285 0.01881 fixed
birth_order_nonlinear2 0.002136 0.005222 0.409 -0.008099 0.01237 fixed
birth_order_nonlinear3 -0.001164 0.006104 -0.1907 -0.01313 0.0108 fixed
birth_order_nonlinear4 0.0004516 0.007227 0.06249 -0.01371 0.01462 fixed
birth_order_nonlinear5 0.006838 0.008807 0.7765 -0.01042 0.0241 fixed
birth_order_nonlinear5+ 0.01633 0.008275 1.974 0.0001131 0.03255 fixed
sd_(Intercept).mother_pidlink 0.05469 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1091 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.01009 0.07847 0.1286 -0.1437 0.1639 fixed
poly(age, 3, raw = TRUE)1 0.002548 0.008429 0.3022 -0.01397 0.01907 fixed
poly(age, 3, raw = TRUE)2 -0.0001574 0.0002892 -0.5441 -0.0007243 0.0004095 fixed
poly(age, 3, raw = TRUE)3 0.000002441 0.000003177 0.7684 -0.000003785 0.000008668 fixed
male 0.003936 0.003938 0.9995 -0.003783 0.01166 fixed
count_birth_order2/2 -0.007212 0.01106 -0.6522 -0.02889 0.01446 fixed
count_birth_order1/3 -0.008172 0.009511 -0.8592 -0.02681 0.01047 fixed
count_birth_order2/3 -0.009063 0.01034 -0.8768 -0.02932 0.0112 fixed
count_birth_order3/3 0.001455 0.01135 0.1282 -0.0208 0.02371 fixed
count_birth_order1/4 0.0004717 0.01034 0.04564 -0.01979 0.02073 fixed
count_birth_order2/4 -0.007483 0.01077 -0.6945 -0.0286 0.01363 fixed
count_birth_order3/4 -0.01528 0.01176 -1.298 -0.03833 0.007783 fixed
count_birth_order4/4 -0.001979 0.01256 -0.1576 -0.02659 0.02263 fixed
count_birth_order1/5 -0.01023 0.01238 -0.8266 -0.03449 0.01403 fixed
count_birth_order2/5 -0.007519 0.01284 -0.5854 -0.03269 0.01765 fixed
count_birth_order3/5 -0.0147 0.01307 -1.125 -0.04031 0.01091 fixed
count_birth_order4/5 0.00125 0.01328 0.09417 -0.02477 0.02728 fixed
count_birth_order5/5 -0.007193 0.01338 -0.5377 -0.03341 0.01902 fixed
count_birth_order1/5+ -0.0104 0.01119 -0.9288 -0.03233 0.01154 fixed
count_birth_order2/5+ 0.02034 0.01181 1.723 -0.002802 0.04349 fixed
count_birth_order3/5+ -0.0001227 0.01137 -0.01078 -0.02242 0.02217 fixed
count_birth_order4/5+ -0.007947 0.01109 -0.7164 -0.02969 0.01379 fixed
count_birth_order5/5+ 0.009955 0.01172 0.8491 -0.01302 0.03293 fixed
count_birth_order5+/5+ 0.01644 0.008933 1.84 -0.001072 0.03395 fixed
sd_(Intercept).mother_pidlink 0.05419 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1092 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -5331 -5262 2676 -5353 NA NA NA
12 -5332 -5256 2678 -5356 2.611 1 0.1061
16 -5326 -5226 2679 -5358 2.522 4 0.6408
26 -5318 -5155 2685 -5370 11.59 10 0.3135

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.001325 0.07952 -0.01667 -0.1572 0.1545 fixed
poly(age, 3, raw = TRUE)1 0.003218 0.008549 0.3765 -0.01354 0.01997 fixed
poly(age, 3, raw = TRUE)2 -0.0001742 0.0002935 -0.5934 -0.0007494 0.000401 fixed
poly(age, 3, raw = TRUE)3 0.00000248 0.000003225 0.7689 -0.000003842 0.000008801 fixed
male 0.003577 0.003991 0.8963 -0.004245 0.0114 fixed
sibling_count3 -0.003884 0.006553 -0.5927 -0.01673 0.008959 fixed
sibling_count4 0.002849 0.006893 0.4133 -0.01066 0.01636 fixed
sibling_count5 -0.00005811 0.007916 -0.007341 -0.01557 0.01546 fixed
sibling_count5+ 0.01903 0.006764 2.814 0.005776 0.03229 fixed
sd_(Intercept).mother_pidlink 0.05496 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1093 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.0002167 0.07955 -0.002724 -0.1561 0.1557 fixed
birth_order 0.000822 0.001276 0.644 -0.00168 0.003324 fixed
poly(age, 3, raw = TRUE)1 0.002991 0.008557 0.3496 -0.01378 0.01976 fixed
poly(age, 3, raw = TRUE)2 -0.0001681 0.0002936 -0.5724 -0.0007436 0.0004075 fixed
poly(age, 3, raw = TRUE)3 0.00000245 0.000003226 0.7596 -0.000003872 0.000008773 fixed
male 0.003571 0.003991 0.8946 -0.004252 0.01139 fixed
sibling_count3 -0.004323 0.006588 -0.6561 -0.01724 0.00859 fixed
sibling_count4 0.001934 0.007037 0.2748 -0.01186 0.01573 fixed
sibling_count5 -0.001542 0.008244 -0.187 -0.0177 0.01462 fixed
sibling_count5+ 0.01604 0.008204 1.955 -0.00003848 0.03212 fixed
sd_(Intercept).mother_pidlink 0.05492 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1093 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.004456 0.07962 0.05596 -0.1516 0.1605 fixed
poly(age, 3, raw = TRUE)1 0.002493 0.008558 0.2913 -0.01428 0.01927 fixed
poly(age, 3, raw = TRUE)2 -0.0001531 0.0002937 -0.5212 -0.0007288 0.0004226 fixed
poly(age, 3, raw = TRUE)3 0.000002327 0.000003228 0.721 -0.000003999 0.000008653 fixed
male 0.00353 0.003992 0.8844 -0.004294 0.01135 fixed
sibling_count3 -0.003619 0.006713 -0.5391 -0.01678 0.009538 fixed
sibling_count4 0.001897 0.007268 0.261 -0.01235 0.01614 fixed
sibling_count5 -0.0000944 0.008571 -0.01101 -0.01689 0.01671 fixed
sibling_count5+ 0.01416 0.008412 1.684 -0.002323 0.03065 fixed
birth_order_nonlinear2 0.003759 0.00514 0.7312 -0.006316 0.01383 fixed
birth_order_nonlinear3 -0.001207 0.00609 -0.1982 -0.01314 0.01073 fixed
birth_order_nonlinear4 0.006397 0.007642 0.837 -0.008582 0.02138 fixed
birth_order_nonlinear5 -0.004586 0.009411 -0.4873 -0.02303 0.01386 fixed
birth_order_nonlinear5+ 0.01514 0.009452 1.602 -0.003388 0.03366 fixed
sd_(Intercept).mother_pidlink 0.05481 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1094 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.001525 0.07984 0.0191 -0.155 0.158 fixed
poly(age, 3, raw = TRUE)1 0.003193 0.008584 0.372 -0.01363 0.02002 fixed
poly(age, 3, raw = TRUE)2 -0.0001795 0.0002947 -0.6092 -0.0007572 0.0003981 fixed
poly(age, 3, raw = TRUE)3 0.000002647 0.000003239 0.817 -0.000003703 0.000008996 fixed
male 0.003806 0.003996 0.9523 -0.004027 0.01164 fixed
count_birth_order2/2 -0.00583 0.009763 -0.5971 -0.02497 0.01331 fixed
count_birth_order1/3 -0.008914 0.008482 -1.051 -0.02554 0.00771 fixed
count_birth_order2/3 -0.006944 0.009415 -0.7375 -0.0254 0.01151 fixed
count_birth_order3/3 0.001039 0.01007 0.1032 -0.0187 0.02078 fixed
count_birth_order1/4 0.005042 0.009935 0.5075 -0.01443 0.02451 fixed
count_birth_order2/4 0.001827 0.01037 0.1762 -0.01849 0.02215 fixed
count_birth_order3/4 -0.01081 0.01086 -0.9954 -0.03209 0.01047 fixed
count_birth_order4/4 0.005862 0.01171 0.5008 -0.01708 0.0288 fixed
count_birth_order1/5 -0.005219 0.01298 -0.4021 -0.03066 0.02022 fixed
count_birth_order2/5 -0.001872 0.01453 -0.1289 -0.03035 0.0266 fixed
count_birth_order3/5 -0.009697 0.01392 -0.6966 -0.03698 0.01758 fixed
count_birth_order4/5 0.01645 0.01343 1.225 -0.00987 0.04277 fixed
count_birth_order5/5 -0.01244 0.01413 -0.8801 -0.04013 0.01526 fixed
count_birth_order1/5+ -0.003465 0.01264 -0.2742 -0.02823 0.0213 fixed
count_birth_order2/5+ 0.03849 0.01278 3.012 0.01345 0.06353 fixed
count_birth_order3/5+ 0.01002 0.01223 0.8194 -0.01395 0.034 fixed
count_birth_order4/5+ 0.006257 0.01216 0.5146 -0.01758 0.03009 fixed
count_birth_order5/5+ 0.009081 0.01151 0.7888 -0.01348 0.03165 fixed
count_birth_order5+/5+ 0.02622 0.009074 2.889 0.008431 0.044 fixed
sd_(Intercept).mother_pidlink 0.05436 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1095 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -5181 -5113 2602 -5203 NA NA NA
12 -5180 -5105 2602 -5204 0.4171 1 0.5184
16 -5177 -5077 2604 -5209 4.736 4 0.3155
26 -5171 -5009 2611 -5223 14.04 10 0.171

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Category_Casual worker not in agriculture

birthorder <- birthorder %>% mutate(outcome = `Category_Casual worker not in agriculture`)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.2572 0.05886 4.37 0.1418 0.3725 fixed
poly(age, 3, raw = TRUE)1 -0.01623 0.00528 -3.074 -0.02658 -0.005884 fixed
poly(age, 3, raw = TRUE)2 0.0003827 0.0001483 2.581 0.00009208 0.0006733 fixed
poly(age, 3, raw = TRUE)3 -0.000002755 0.000001317 -2.092 -0.000005337 -0.0000001736 fixed
male 0.06204 0.00543 11.43 0.0514 0.07269 fixed
sibling_count3 0.002887 0.01133 0.2549 -0.01932 0.02509 fixed
sibling_count4 -0.006908 0.01143 -0.6044 -0.02931 0.01549 fixed
sibling_count5 -0.001617 0.01191 -0.1358 -0.02496 0.02173 fixed
sibling_count5+ -0.007001 0.009233 -0.7582 -0.0251 0.0111 fixed
sd_(Intercept).mother_pidlink 0.09542 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2537 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.2572 0.05886 4.37 0.1419 0.3726 fixed
birth_order -0.0001545 0.001106 -0.1397 -0.002322 0.002013 fixed
poly(age, 3, raw = TRUE)1 -0.01619 0.005289 -3.061 -0.02656 -0.005824 fixed
poly(age, 3, raw = TRUE)2 0.0003811 0.0001487 2.563 0.00008966 0.0006725 fixed
poly(age, 3, raw = TRUE)3 -0.000002741 0.000001321 -2.075 -0.000005331 -0.0000001514 fixed
male 0.06205 0.00543 11.43 0.0514 0.07269 fixed
sibling_count3 0.002928 0.01133 0.2583 -0.01928 0.02514 fixed
sibling_count4 -0.006814 0.01145 -0.5951 -0.02925 0.01563 fixed
sibling_count5 -0.00145 0.01197 -0.1212 -0.02491 0.02201 fixed
sibling_count5+ -0.00646 0.01001 -0.6453 -0.02608 0.01316 fixed
sd_(Intercept).mother_pidlink 0.09546 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2537 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.2574 0.05896 4.367 0.1419 0.373 fixed
poly(age, 3, raw = TRUE)1 -0.0162 0.005294 -3.06 -0.02657 -0.005822 fixed
poly(age, 3, raw = TRUE)2 0.0003819 0.0001488 2.567 0.00009029 0.0006736 fixed
poly(age, 3, raw = TRUE)3 -0.000002754 0.000001322 -2.083 -0.000005345 -0.0000001623 fixed
male 0.06205 0.005432 11.42 0.05141 0.0727 fixed
sibling_count3 0.003057 0.0115 0.2658 -0.01948 0.0256 fixed
sibling_count4 -0.006794 0.01176 -0.578 -0.02984 0.01625 fixed
sibling_count5 -0.0008552 0.01238 -0.0691 -0.02512 0.0234 fixed
sibling_count5+ -0.006345 0.01048 -0.6053 -0.02689 0.0142 fixed
birth_order_nonlinear2 -0.00146 0.007906 -0.1847 -0.01696 0.01404 fixed
birth_order_nonlinear3 -0.001161 0.009188 -0.1263 -0.01917 0.01685 fixed
birth_order_nonlinear4 -0.0003736 0.01032 -0.03621 -0.0206 0.01985 fixed
birth_order_nonlinear5 -0.004592 0.01161 -0.3955 -0.02734 0.01816 fixed
birth_order_nonlinear5+ -0.0009382 0.009655 -0.09716 -0.01986 0.01799 fixed
sd_(Intercept).mother_pidlink 0.09536 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2537 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.2538 0.05915 4.291 0.1379 0.3697 fixed
poly(age, 3, raw = TRUE)1 -0.01604 0.005297 -3.028 -0.02642 -0.00566 fixed
poly(age, 3, raw = TRUE)2 0.000377 0.0001488 2.533 0.00008526 0.0006687 fixed
poly(age, 3, raw = TRUE)3 -0.000002706 0.000001322 -2.046 -0.000005298 -0.0000001144 fixed
male 0.06207 0.005433 11.42 0.05142 0.07272 fixed
count_birth_order2/2 0.004305 0.0157 0.2742 -0.02646 0.03507 fixed
count_birth_order1/3 0.008432 0.01522 0.5541 -0.02139 0.03826 fixed
count_birth_order2/3 0.00825 0.01686 0.4892 -0.0248 0.0413 fixed
count_birth_order3/3 -0.008015 0.01852 -0.4328 -0.04431 0.02828 fixed
count_birth_order1/4 -0.0248 0.01667 -1.488 -0.05747 0.007867 fixed
count_birth_order2/4 0.007322 0.01787 0.4098 -0.0277 0.04235 fixed
count_birth_order3/4 0.01171 0.01905 0.6149 -0.02562 0.04905 fixed
count_birth_order4/4 -0.009058 0.02023 -0.4477 -0.04872 0.0306 fixed
count_birth_order1/5 -0.01351 0.01892 -0.7138 -0.05059 0.02358 fixed
count_birth_order2/5 -0.01094 0.02006 -0.5453 -0.05026 0.02838 fixed
count_birth_order3/5 0.02642 0.02112 1.251 -0.01497 0.06782 fixed
count_birth_order4/5 0.01116 0.0222 0.5028 -0.03235 0.05466 fixed
count_birth_order5/5 -0.006309 0.02219 -0.2843 -0.04981 0.03719 fixed
count_birth_order1/5+ 0.01452 0.01457 0.9961 -0.01405 0.04308 fixed
count_birth_order2/5+ -0.01502 0.01518 -0.9897 -0.04477 0.01473 fixed
count_birth_order3/5+ -0.01649 0.01495 -1.103 -0.04579 0.01281 fixed
count_birth_order4/5+ -0.005982 0.01468 -0.4074 -0.03476 0.0228 fixed
count_birth_order5/5+ -0.007781 0.01485 -0.524 -0.03688 0.02132 fixed
count_birth_order5+/5+ -0.005229 0.01203 -0.4346 -0.02881 0.01835 fixed
sd_(Intercept).mother_pidlink 0.09533 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2537 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 2203 2282 -1090 2181 NA NA NA
12 2205 2291 -1090 2181 0.01918 1 0.8899
16 2213 2328 -1090 2181 0.1667 4 0.9967
26 2218 2406 -1083 2166 14.27 10 0.1609

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.5278 0.1723 3.064 0.1902 0.8654 fixed
poly(age, 3, raw = TRUE)1 -0.04411 0.0185 -2.385 -0.08037 -0.007856 fixed
poly(age, 3, raw = TRUE)2 0.00134 0.0006347 2.111 0.00009561 0.002583 fixed
poly(age, 3, raw = TRUE)3 -0.00001395 0.000006968 -2.001 -0.0000276 -0.0000002891 fixed
male 0.04766 0.008681 5.49 0.03064 0.06467 fixed
sibling_count3 -0.005651 0.01432 -0.3946 -0.03372 0.02242 fixed
sibling_count4 0.009222 0.0148 0.6229 -0.01979 0.03824 fixed
sibling_count5 -0.01576 0.01645 -0.9578 -0.04801 0.01649 fixed
sibling_count5+ 0.02188 0.01429 1.531 -0.006124 0.04988 fixed
sd_(Intercept).mother_pidlink 0.08695 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2512 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.5268 0.1723 3.058 0.1891 0.8645 fixed
birth_order -0.00116 0.002699 -0.4297 -0.006449 0.00413 fixed
poly(age, 3, raw = TRUE)1 -0.04386 0.01851 -2.37 -0.08014 -0.007585 fixed
poly(age, 3, raw = TRUE)2 0.001333 0.0006349 2.1 0.00008914 0.002578 fixed
poly(age, 3, raw = TRUE)3 -0.00001393 0.000006969 -1.999 -0.00002759 -0.0000002715 fixed
male 0.04769 0.008682 5.493 0.03067 0.06471 fixed
sibling_count3 -0.005053 0.01439 -0.3512 -0.03326 0.02315 fixed
sibling_count4 0.01051 0.01511 0.6959 -0.0191 0.04012 fixed
sibling_count5 -0.01363 0.01718 -0.7934 -0.04731 0.02005 fixed
sibling_count5+ 0.02605 0.01727 1.508 -0.007805 0.05991 fixed
sd_(Intercept).mother_pidlink 0.08704 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2512 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.5254 0.1725 3.046 0.1873 0.8635 fixed
poly(age, 3, raw = TRUE)1 -0.0439 0.01852 -2.37 -0.0802 -0.007599 fixed
poly(age, 3, raw = TRUE)2 0.001338 0.0006353 2.106 0.00009249 0.002583 fixed
poly(age, 3, raw = TRUE)3 -0.00001404 0.000006975 -2.013 -0.00002771 -0.0000003718 fixed
male 0.04763 0.008685 5.484 0.03061 0.06465 fixed
sibling_count3 -0.001229 0.01468 -0.08376 -0.03 0.02754 fixed
sibling_count4 0.01619 0.01565 1.035 -0.01448 0.04687 fixed
sibling_count5 -0.00553 0.01801 -0.3071 -0.04082 0.02976 fixed
sibling_count5+ 0.03521 0.01774 1.985 0.0004435 0.06997 fixed
birth_order_nonlinear2 0.0003386 0.01147 0.02952 -0.02214 0.02282 fixed
birth_order_nonlinear3 -0.01783 0.01349 -1.322 -0.04427 0.008611 fixed
birth_order_nonlinear4 -0.01307 0.01643 -0.7954 -0.04528 0.01914 fixed
birth_order_nonlinear5 -0.01722 0.02021 -0.8518 -0.05683 0.02239 fixed
birth_order_nonlinear5+ -0.01704 0.02005 -0.8501 -0.05634 0.02225 fixed
sd_(Intercept).mother_pidlink 0.08715 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2512 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.5537 0.1731 3.199 0.2145 0.8929 fixed
poly(age, 3, raw = TRUE)1 -0.04629 0.01858 -2.491 -0.08271 -0.009864 fixed
poly(age, 3, raw = TRUE)2 0.001425 0.0006377 2.234 0.000175 0.002675 fixed
poly(age, 3, raw = TRUE)3 -0.00001505 0.000007003 -2.15 -0.00002878 -0.00000133 fixed
male 0.04776 0.008695 5.492 0.03071 0.0648 fixed
count_birth_order2/2 -0.02316 0.02227 -1.04 -0.06681 0.02049 fixed
count_birth_order1/3 -0.01659 0.01882 -0.8817 -0.05348 0.02029 fixed
count_birth_order2/3 0.002044 0.02091 0.09778 -0.03893 0.04302 fixed
count_birth_order3/3 -0.0255 0.02262 -1.127 -0.06982 0.01883 fixed
count_birth_order1/4 -0.0171 0.02162 -0.7912 -0.05947 0.02527 fixed
count_birth_order2/4 0.01951 0.02274 0.858 -0.02506 0.06407 fixed
count_birth_order3/4 0.004401 0.02385 0.1845 -0.04235 0.05115 fixed
count_birth_order4/4 0.006849 0.0253 0.2707 -0.04274 0.05644 fixed
count_birth_order1/5 -0.0173 0.02837 -0.6099 -0.0729 0.0383 fixed
count_birth_order2/5 -0.008308 0.03089 -0.269 -0.06884 0.05223 fixed
count_birth_order3/5 -0.03295 0.02893 -1.139 -0.08966 0.02376 fixed
count_birth_order4/5 -0.03155 0.02815 -1.121 -0.08671 0.02362 fixed
count_birth_order5/5 -0.02164 0.02952 -0.7332 -0.07949 0.03621 fixed
count_birth_order1/5+ 0.07584 0.02698 2.811 0.02297 0.1287 fixed
count_birth_order2/5+ 0.01185 0.02723 0.4352 -0.04152 0.06523 fixed
count_birth_order3/5+ -0.007979 0.02663 -0.2996 -0.06017 0.04421 fixed
count_birth_order4/5+ 0.007473 0.02587 0.2889 -0.04322 0.05817 fixed
count_birth_order5/5+ 0.004793 0.02494 0.1922 -0.04409 0.05368 fixed
count_birth_order5+/5+ 0.01042 0.01941 0.5368 -0.02762 0.04846 fixed
sd_(Intercept).mother_pidlink 0.08705 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2512 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 715.1 783.9 -346.6 693.1 NA NA NA
12 716.9 791.9 -346.5 692.9 0.1845 1 0.6675
16 722.4 822.4 -345.2 690.4 2.501 4 0.6444
26 731 893.5 -339.5 679 11.45 10 0.3232

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.5216 0.1714 3.044 0.1858 0.8575 fixed
poly(age, 3, raw = TRUE)1 -0.04348 0.01842 -2.36 -0.07958 -0.007375 fixed
poly(age, 3, raw = TRUE)2 0.001321 0.0006321 2.09 0.00008216 0.00256 fixed
poly(age, 3, raw = TRUE)3 -0.00001375 0.000006941 -1.98 -0.00002735 -0.0000001409 fixed
male 0.04835 0.008643 5.595 0.03141 0.0653 fixed
sibling_count3 -0.01324 0.01568 -0.8443 -0.04397 0.01749 fixed
sibling_count4 -0.005395 0.01581 -0.3412 -0.03638 0.02559 fixed
sibling_count5 0.008151 0.01666 0.4892 -0.02451 0.04081 fixed
sibling_count5+ 0.01342 0.01461 0.9192 -0.0152 0.04205 fixed
sd_(Intercept).mother_pidlink 0.08703 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2509 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.5211 0.1714 3.041 0.1852 0.857 fixed
birth_order -0.0006892 0.002377 -0.2899 -0.005349 0.00397 fixed
poly(age, 3, raw = TRUE)1 -0.04334 0.01843 -2.352 -0.07946 -0.007223 fixed
poly(age, 3, raw = TRUE)2 0.001318 0.0006322 2.084 0.00007869 0.002557 fixed
poly(age, 3, raw = TRUE)3 -0.00001374 0.000006942 -1.979 -0.00002735 -0.0000001349 fixed
male 0.04837 0.008644 5.596 0.03143 0.06531 fixed
sibling_count3 -0.01288 0.01573 -0.8184 -0.04371 0.01796 fixed
sibling_count4 -0.004677 0.016 -0.2923 -0.03605 0.02669 fixed
sibling_count5 0.00933 0.01715 0.544 -0.02429 0.04295 fixed
sibling_count5+ 0.01586 0.01685 0.9413 -0.01716 0.04888 fixed
sd_(Intercept).mother_pidlink 0.0871 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2509 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.5174 0.1715 3.017 0.1813 0.8536 fixed
poly(age, 3, raw = TRUE)1 -0.04304 0.01843 -2.335 -0.07916 -0.006915 fixed
poly(age, 3, raw = TRUE)2 0.001309 0.0006324 2.069 0.00006903 0.002548 fixed
poly(age, 3, raw = TRUE)3 -0.0000137 0.000006945 -1.973 -0.00002731 -0.00000008901 fixed
male 0.04827 0.008643 5.585 0.03133 0.06521 fixed
sibling_count3 -0.008033 0.01601 -0.5018 -0.03941 0.02334 fixed
sibling_count4 0.004259 0.01648 0.2584 -0.02804 0.03656 fixed
sibling_count5 0.01848 0.01791 1.032 -0.01662 0.05357 fixed
sibling_count5+ 0.02776 0.01734 1.602 -0.006214 0.06174 fixed
birth_order_nonlinear2 0.001952 0.0116 0.1683 -0.02078 0.02469 fixed
birth_order_nonlinear3 -0.02108 0.01354 -1.558 -0.04761 0.005446 fixed
birth_order_nonlinear4 -0.0265 0.01601 -1.655 -0.05787 0.004873 fixed
birth_order_nonlinear5 -0.002944 0.0195 -0.151 -0.04116 0.03527 fixed
birth_order_nonlinear5+ -0.01788 0.01808 -0.989 -0.05332 0.01756 fixed
sd_(Intercept).mother_pidlink 0.08779 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2506 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.5206 0.1719 3.028 0.1836 0.8575 fixed
poly(age, 3, raw = TRUE)1 -0.04292 0.01848 -2.323 -0.07913 -0.006707 fixed
poly(age, 3, raw = TRUE)2 0.00131 0.0006341 2.066 0.00006733 0.002553 fixed
poly(age, 3, raw = TRUE)3 -0.00001377 0.000006966 -1.977 -0.00002742 -0.000000116 fixed
male 0.04851 0.008658 5.603 0.03154 0.06548 fixed
count_birth_order2/2 -0.0166 0.02448 -0.6779 -0.06458 0.03138 fixed
count_birth_order1/3 -0.02012 0.02077 -0.9688 -0.06082 0.02058 fixed
count_birth_order2/3 -0.003152 0.0226 -0.1394 -0.04745 0.04115 fixed
count_birth_order3/3 -0.03641 0.02486 -1.465 -0.08513 0.01231 fixed
count_birth_order1/4 -0.02499 0.02258 -1.107 -0.06926 0.01927 fixed
count_birth_order2/4 0.008599 0.02357 0.3648 -0.03761 0.0548 fixed
count_birth_order3/4 -0.01663 0.02578 -0.6451 -0.06717 0.0339 fixed
count_birth_order4/4 -0.006556 0.02753 -0.2381 -0.06052 0.0474 fixed
count_birth_order1/5 0.01083 0.02708 0.3998 -0.04225 0.06391 fixed
count_birth_order2/5 0.01599 0.02815 0.5683 -0.03917 0.07116 fixed
count_birth_order3/5 0.01195 0.02866 0.417 -0.04422 0.06812 fixed
count_birth_order4/5 -0.01794 0.02915 -0.6154 -0.07506 0.03919 fixed
count_birth_order5/5 -0.01009 0.02936 -0.3437 -0.06764 0.04745 fixed
count_birth_order1/5+ 0.04959 0.02449 2.025 0.001597 0.09759 fixed
count_birth_order2/5+ 0.01055 0.0259 0.4072 -0.04022 0.06132 fixed
count_birth_order3/5+ -0.01746 0.02495 -0.6998 -0.06636 0.03144 fixed
count_birth_order4/5+ -0.01776 0.02434 -0.7295 -0.06546 0.02995 fixed
count_birth_order5/5+ 0.03147 0.02575 1.222 -0.019 0.08195 fixed
count_birth_order5+/5+ 0.00355 0.01941 0.1829 -0.03448 0.04158 fixed
sd_(Intercept).mother_pidlink 0.08629 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2511 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 711.8 780.7 -344.9 689.8 NA NA NA
12 713.7 788.8 -344.9 689.7 0.08395 1 0.772
16 716.2 816.4 -342.1 684.2 5.493 4 0.2404
26 726.6 889.3 -337.3 674.6 9.669 10 0.47

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.5159 0.1741 2.964 0.1747 0.857 fixed
poly(age, 3, raw = TRUE)1 -0.04311 0.01872 -2.303 -0.07981 -0.006419 fixed
poly(age, 3, raw = TRUE)2 0.001318 0.0006429 2.05 0.000058 0.002578 fixed
poly(age, 3, raw = TRUE)3 -0.00001381 0.000007066 -1.955 -0.00002766 0.00000003501 fixed
male 0.04962 0.008764 5.662 0.03244 0.0668 fixed
sibling_count3 -0.007423 0.01403 -0.5289 -0.03493 0.02008 fixed
sibling_count4 0.003436 0.01471 0.2335 -0.0254 0.03227 fixed
sibling_count5 -0.02066 0.01679 -1.23 -0.05357 0.01225 fixed
sibling_count5+ 0.02131 0.01435 1.485 -0.006809 0.04943 fixed
sd_(Intercept).mother_pidlink 0.08655 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2514 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.5148 0.1741 2.956 0.1735 0.856 fixed
birth_order -0.0009926 0.002778 -0.3573 -0.006437 0.004452 fixed
poly(age, 3, raw = TRUE)1 -0.04287 0.01874 -2.288 -0.07959 -0.006147 fixed
poly(age, 3, raw = TRUE)2 0.001312 0.0006432 2.04 0.00005121 0.002572 fixed
poly(age, 3, raw = TRUE)3 -0.00001379 0.000007067 -1.951 -0.00002764 0.00000006217 fixed
male 0.04963 0.008765 5.662 0.03245 0.0668 fixed
sibling_count3 -0.006901 0.01411 -0.4891 -0.03456 0.02076 fixed
sibling_count4 0.004525 0.01502 0.3012 -0.02492 0.03397 fixed
sibling_count5 -0.01889 0.01751 -1.079 -0.0532 0.01542 fixed
sibling_count5+ 0.02488 0.01747 1.424 -0.009374 0.05913 fixed
sd_(Intercept).mother_pidlink 0.08661 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2514 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.5132 0.1743 2.944 0.1715 0.8548 fixed
poly(age, 3, raw = TRUE)1 -0.04284 0.01875 -2.285 -0.07958 -0.006096 fixed
poly(age, 3, raw = TRUE)2 0.001315 0.0006437 2.044 0.00005383 0.002577 fixed
poly(age, 3, raw = TRUE)3 -0.00001392 0.000007074 -1.968 -0.00002778 -0.0000000542 fixed
male 0.04953 0.00877 5.648 0.03234 0.06672 fixed
sibling_count3 -0.004327 0.01441 -0.3003 -0.03257 0.02392 fixed
sibling_count4 0.009017 0.01557 0.5791 -0.0215 0.03954 fixed
sibling_count5 -0.01119 0.01828 -0.6119 -0.04702 0.02465 fixed
sibling_count5+ 0.0347 0.01797 1.931 -0.0005219 0.06991 fixed
birth_order_nonlinear2 -0.001036 0.01143 -0.09066 -0.02344 0.02136 fixed
birth_order_nonlinear3 -0.01195 0.01352 -0.8841 -0.03845 0.01454 fixed
birth_order_nonlinear4 -0.01326 0.01694 -0.7827 -0.04646 0.01994 fixed
birth_order_nonlinear5 -0.02167 0.02087 -1.038 -0.06258 0.01924 fixed
birth_order_nonlinear5+ -0.01784 0.02069 -0.8624 -0.05839 0.02271 fixed
sd_(Intercept).mother_pidlink 0.08642 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2516 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.5353 0.1748 3.062 0.1927 0.8779 fixed
poly(age, 3, raw = TRUE)1 -0.04495 0.0188 -2.391 -0.0818 -0.008104 fixed
poly(age, 3, raw = TRUE)2 0.001394 0.0006457 2.158 0.0001281 0.002659 fixed
poly(age, 3, raw = TRUE)3 -0.00001485 0.000007099 -2.092 -0.00002876 -0.0000009356 fixed
male 0.04938 0.00878 5.625 0.03218 0.06659 fixed
count_birth_order2/2 -0.0126 0.02165 -0.5819 -0.05504 0.02984 fixed
count_birth_order1/3 -0.0158 0.01851 -0.8535 -0.05207 0.02048 fixed
count_birth_order2/3 0.0005413 0.02059 0.02629 -0.03981 0.04089 fixed
count_birth_order3/3 -0.01806 0.02204 -0.8195 -0.06126 0.02514 fixed
count_birth_order1/4 -0.01388 0.02171 -0.6392 -0.05642 0.02867 fixed
count_birth_order2/4 0.01046 0.02269 0.461 -0.03401 0.05493 fixed
count_birth_order3/4 0.00654 0.02379 0.275 -0.04008 0.05316 fixed
count_birth_order4/4 -0.001575 0.02566 -0.06138 -0.05186 0.04871 fixed
count_birth_order1/5 -0.02303 0.0284 -0.811 -0.07869 0.03263 fixed
count_birth_order2/5 0.007691 0.03189 0.2412 -0.0548 0.07019 fixed
count_birth_order3/5 -0.03383 0.03054 -1.108 -0.09369 0.02603 fixed
count_birth_order4/5 -0.04271 0.02946 -1.45 -0.1004 0.01502 fixed
count_birth_order5/5 -0.02514 0.03101 -0.8108 -0.08592 0.03564 fixed
count_birth_order1/5+ 0.08574 0.02768 3.098 0.0315 0.14 fixed
count_birth_order2/5+ -0.0053 0.02803 -0.189 -0.06025 0.04965 fixed
count_birth_order3/5+ 0.003049 0.02684 0.1136 -0.04956 0.05566 fixed
count_birth_order4/5+ 0.02215 0.02671 0.8294 -0.0302 0.0745 fixed
count_birth_order5/5+ 0.00261 0.02527 0.1033 -0.04692 0.05214 fixed
count_birth_order5+/5+ 0.01287 0.01968 0.6538 -0.0257 0.05144 fixed
sd_(Intercept).mother_pidlink 0.08667 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2514 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 705.5 774 -341.7 683.5 NA NA NA
12 707.4 782.2 -341.7 683.4 0.1277 1 0.7209
16 713.6 813.3 -340.8 681.6 1.778 4 0.7765
26 721 883.1 -334.5 669 12.58 10 0.2478

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Category_Government worker

birthorder <- birthorder %>% mutate(outcome = `Category_Government worker`)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.1944 0.05324 -3.651 -0.2987 -0.09004 fixed
poly(age, 3, raw = TRUE)1 0.01928 0.004779 4.034 0.009912 0.02865 fixed
poly(age, 3, raw = TRUE)2 -0.0004177 0.0001344 -3.108 -0.0006812 -0.0001543 fixed
poly(age, 3, raw = TRUE)3 0.000002971 0.000001196 2.484 0.0000006272 0.000005316 fixed
male -0.02151 0.004853 -4.432 -0.03102 -0.012 fixed
sibling_count3 0.008502 0.01049 0.8108 -0.01205 0.02905 fixed
sibling_count4 0.01057 0.0106 0.9964 -0.01022 0.03135 fixed
sibling_count5 0.01661 0.01108 1.498 -0.005117 0.03833 fixed
sibling_count5+ -0.009424 0.008552 -1.102 -0.02619 0.007337 fixed
sd_(Intercept).mother_pidlink 0.1135 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2183 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.1943 0.05324 -3.649 -0.2986 -0.08993 fixed
birth_order -0.0002719 0.001006 -0.2703 -0.002244 0.0017 fixed
poly(age, 3, raw = TRUE)1 0.01936 0.004788 4.043 0.009972 0.02874 fixed
poly(age, 3, raw = TRUE)2 -0.0004207 0.0001349 -3.119 -0.000685 -0.0001563 fixed
poly(age, 3, raw = TRUE)3 0.000002999 0.0000012 2.498 0.000000646 0.000005352 fixed
male -0.02151 0.004854 -4.431 -0.03102 -0.01199 fixed
sibling_count3 0.008566 0.01049 0.8168 -0.01199 0.02912 fixed
sibling_count4 0.01072 0.01062 1.009 -0.01009 0.03153 fixed
sibling_count5 0.01689 0.01113 1.517 -0.004931 0.0387 fixed
sibling_count5+ -0.008494 0.00922 -0.9212 -0.02656 0.009577 fixed
sd_(Intercept).mother_pidlink 0.1135 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2183 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.1934 0.0533 -3.628 -0.2979 -0.08892 fixed
poly(age, 3, raw = TRUE)1 0.01971 0.004791 4.115 0.01032 0.0291 fixed
poly(age, 3, raw = TRUE)2 -0.0004282 0.0001349 -3.175 -0.0006926 -0.0001639 fixed
poly(age, 3, raw = TRUE)3 0.000003015 0.000001201 2.511 0.0000006618 0.000005368 fixed
male -0.02147 0.004854 -4.424 -0.03098 -0.01196 fixed
sibling_count3 0.01191 0.01062 1.121 -0.008906 0.03273 fixed
sibling_count4 0.01411 0.01086 1.299 -0.007181 0.03541 fixed
sibling_count5 0.02163 0.01146 1.888 -0.0008248 0.04409 fixed
sibling_count5+ -0.002547 0.009608 -0.2651 -0.02138 0.01628 fixed
birth_order_nonlinear2 -0.01229 0.007011 -1.753 -0.02603 0.001451 fixed
birth_order_nonlinear3 -0.01874 0.008133 -2.304 -0.03468 -0.002798 fixed
birth_order_nonlinear4 -0.00766 0.009138 -0.8382 -0.02557 0.01025 fixed
birth_order_nonlinear5 -0.01681 0.01028 -1.634 -0.03696 0.003347 fixed
birth_order_nonlinear5+ -0.0144 0.008664 -1.662 -0.03138 0.002579 fixed
sd_(Intercept).mother_pidlink 0.1133 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2184 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.1926 0.05348 -3.602 -0.2975 -0.08783 fixed
poly(age, 3, raw = TRUE)1 0.0199 0.004794 4.151 0.0105 0.0293 fixed
poly(age, 3, raw = TRUE)2 -0.0004295 0.000135 -3.182 -0.000694 -0.000165 fixed
poly(age, 3, raw = TRUE)3 0.000002987 0.000001201 2.486 0.0000006324 0.000005341 fixed
male -0.02117 0.004856 -4.36 -0.03069 -0.01165 fixed
count_birth_order2/2 -0.0235 0.01393 -1.687 -0.0508 0.003795 fixed
count_birth_order1/3 0.02031 0.01377 1.475 -0.00668 0.04729 fixed
count_birth_order2/3 -0.01708 0.01523 -1.121 -0.04693 0.01278 fixed
count_birth_order3/3 -0.01838 0.0167 -1.101 -0.05112 0.01435 fixed
count_birth_order1/4 0.009319 0.01506 0.6189 -0.02019 0.03883 fixed
count_birth_order2/4 0.01105 0.01613 0.685 -0.02056 0.04265 fixed
count_birth_order3/4 -0.0106 0.01717 -0.6175 -0.04426 0.02305 fixed
count_birth_order4/4 -0.01459 0.01822 -0.8008 -0.05031 0.02113 fixed
count_birth_order1/5 0.01084 0.01706 0.6352 -0.0226 0.04428 fixed
count_birth_order2/5 -0.002587 0.01808 -0.1431 -0.03802 0.03285 fixed
count_birth_order3/5 0.01388 0.01901 0.7301 -0.02338 0.05115 fixed
count_birth_order4/5 0.006791 0.01995 0.3404 -0.03231 0.04589 fixed
count_birth_order5/5 0.006184 0.01996 0.3098 -0.03294 0.04531 fixed
count_birth_order1/5+ -0.01924 0.01312 -1.466 -0.04496 0.006483 fixed
count_birth_order2/5+ -0.01011 0.01365 -0.7407 -0.03687 0.01664 fixed
count_birth_order3/5+ -0.027 0.01345 -2.008 -0.05336 -0.0006427 fixed
count_birth_order4/5+ -0.007687 0.01322 -0.5815 -0.0336 0.01822 fixed
count_birth_order5/5+ -0.02553 0.01336 -1.911 -0.05171 0.000652 fixed
count_birth_order5+/5+ -0.02128 0.01098 -1.938 -0.0428 0.0002367 fixed
sd_(Intercept).mother_pidlink 0.1134 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2183 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 72.34 151.8 -25.17 50.34 NA NA NA
12 74.27 160.9 -25.13 50.27 0.0738 1 0.7859
16 75.28 190.8 -21.64 43.28 6.987 4 0.1366
26 84.51 272.2 -16.26 32.51 10.77 10 0.3758

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.867 0.1698 -5.107 -1.2 -0.5343 fixed
poly(age, 3, raw = TRUE)1 0.09225 0.01822 5.062 0.05653 0.128 fixed
poly(age, 3, raw = TRUE)2 -0.002883 0.000625 -4.613 -0.004108 -0.001658 fixed
poly(age, 3, raw = TRUE)3 0.00003017 0.000006861 4.397 0.00001672 0.00004362 fixed
male -0.02578 0.008528 -3.023 -0.0425 -0.009067 fixed
sibling_count3 0.007785 0.01443 0.5395 -0.0205 0.03607 fixed
sibling_count4 0.004312 0.01497 0.2881 -0.02502 0.03365 fixed
sibling_count5 0.003921 0.01672 0.2346 -0.02884 0.03668 fixed
sibling_count5+ -0.0358 0.01452 -2.465 -0.06425 -0.007338 fixed
sd_(Intercept).mother_pidlink 0.1191 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2355 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.871 0.1698 -5.13 -1.204 -0.5382 fixed
birth_order -0.003387 0.002677 -1.265 -0.008633 0.00186 fixed
poly(age, 3, raw = TRUE)1 0.09312 0.01824 5.106 0.05738 0.1289 fixed
poly(age, 3, raw = TRUE)2 -0.002906 0.0006253 -4.648 -0.004131 -0.00168 fixed
poly(age, 3, raw = TRUE)3 0.00003026 0.000006861 4.411 0.00001681 0.00004371 fixed
male -0.02566 0.008528 -3.009 -0.04238 -0.00895 fixed
sibling_count3 0.009541 0.01449 0.6583 -0.01887 0.03795 fixed
sibling_count4 0.008137 0.01526 0.5331 -0.02178 0.03805 fixed
sibling_count5 0.0102 0.01743 0.5852 -0.02396 0.04436 fixed
sibling_count5+ -0.02346 0.01749 -1.341 -0.05774 0.01082 fixed
sd_(Intercept).mother_pidlink 0.1187 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2356 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.8585 0.17 -5.05 -1.192 -0.5253 fixed
poly(age, 3, raw = TRUE)1 0.09185 0.01824 5.034 0.05609 0.1276 fixed
poly(age, 3, raw = TRUE)2 -0.002859 0.0006256 -4.571 -0.004086 -0.001633 fixed
poly(age, 3, raw = TRUE)3 0.00002975 0.000006867 4.332 0.00001629 0.00004321 fixed
male -0.02605 0.00853 -3.054 -0.04277 -0.009332 fixed
sibling_count3 0.01138 0.01476 0.7712 -0.01755 0.04031 fixed
sibling_count4 0.007978 0.01577 0.506 -0.02293 0.03888 fixed
sibling_count5 0.01 0.01819 0.5499 -0.02565 0.04566 fixed
sibling_count5+ -0.02489 0.01792 -1.389 -0.06001 0.01024 fixed
birth_order_nonlinear2 -0.01938 0.01113 -1.741 -0.04121 0.002441 fixed
birth_order_nonlinear3 -0.01584 0.01312 -1.207 -0.04154 0.009869 fixed
birth_order_nonlinear4 -0.004245 0.016 -0.2654 -0.0356 0.02711 fixed
birth_order_nonlinear5 -0.02336 0.01968 -1.187 -0.06192 0.01521 fixed
birth_order_nonlinear5+ -0.02327 0.01977 -1.177 -0.06203 0.01548 fixed
sd_(Intercept).mother_pidlink 0.1191 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2355 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.837 0.1707 -4.905 -1.172 -0.5026 fixed
poly(age, 3, raw = TRUE)1 0.08981 0.01832 4.903 0.05391 0.1257 fixed
poly(age, 3, raw = TRUE)2 -0.00279 0.0006283 -4.441 -0.004022 -0.001559 fixed
poly(age, 3, raw = TRUE)3 0.00002901 0.000006898 4.205 0.00001549 0.00004253 fixed
male -0.02647 0.008544 -3.098 -0.04321 -0.00972 fixed
count_birth_order2/2 -0.02663 0.02168 -1.228 -0.06913 0.01586 fixed
count_birth_order1/3 0.0166 0.01863 0.8912 -0.01991 0.05311 fixed
count_birth_order2/3 -0.01654 0.02066 -0.8007 -0.05702 0.02395 fixed
count_birth_order3/3 -0.01347 0.02232 -0.6033 -0.05721 0.03028 fixed
count_birth_order1/4 -0.01372 0.02137 -0.6419 -0.0556 0.02817 fixed
count_birth_order2/4 0.01028 0.02244 0.4582 -0.0337 0.05427 fixed
count_birth_order3/4 0.003243 0.02351 0.1379 -0.04284 0.04932 fixed
count_birth_order4/4 -0.01605 0.02493 -0.644 -0.06492 0.03281 fixed
count_birth_order1/5 0.01175 0.02799 0.4196 -0.04312 0.06661 fixed
count_birth_order2/5 -0.03345 0.03039 -1.101 -0.09303 0.02612 fixed
count_birth_order3/5 -0.03181 0.02849 -1.116 -0.08764 0.02403 fixed
count_birth_order4/5 0.02514 0.02771 0.9072 -0.02917 0.07944 fixed
count_birth_order5/5 -0.000456 0.02904 -0.0157 -0.05738 0.05647 fixed
count_birth_order1/5+ -0.02476 0.0266 -0.9311 -0.07689 0.02737 fixed
count_birth_order2/5+ -0.04955 0.0268 -1.849 -0.1021 0.002986 fixed
count_birth_order3/5+ -0.03274 0.0262 -1.25 -0.08409 0.0186 fixed
count_birth_order4/5+ -0.03021 0.02543 -1.188 -0.08006 0.01963 fixed
count_birth_order5/5+ -0.06004 0.02453 -2.447 -0.1081 -0.01195 fixed
count_birth_order5+/5+ -0.05057 0.01932 -2.617 -0.08844 -0.0127 fixed
sd_(Intercept).mother_pidlink 0.1189 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2356 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 613.9 682.6 -295.9 591.9 NA NA NA
12 614.3 689.3 -295.1 590.3 1.606 1 0.205
16 619.3 719.3 -293.7 587.3 2.942 4 0.5676
26 630.8 793.3 -289.4 578.8 8.559 10 0.5744

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.8577 0.1687 -5.085 -1.188 -0.5271 fixed
poly(age, 3, raw = TRUE)1 0.09078 0.01812 5.01 0.05527 0.1263 fixed
poly(age, 3, raw = TRUE)2 -0.002845 0.0006216 -4.577 -0.004064 -0.001627 fixed
poly(age, 3, raw = TRUE)3 0.00002981 0.000006826 4.367 0.00001643 0.00004319 fixed
male -0.02579 0.008479 -3.041 -0.04241 -0.009168 fixed
sibling_count3 0.01011 0.01575 0.6416 -0.02077 0.04098 fixed
sibling_count4 0.02064 0.01592 1.297 -0.01055 0.05184 fixed
sibling_count5 0.01602 0.01683 0.9516 -0.01697 0.049 fixed
sibling_count5+ -0.01717 0.01473 -1.166 -0.04605 0.0117 fixed
sd_(Intercept).mother_pidlink 0.1183 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2351 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.8617 0.1686 -5.11 -1.192 -0.5311 fixed
birth_order -0.004012 0.00236 -1.7 -0.008638 0.0006143 fixed
poly(age, 3, raw = TRUE)1 0.09173 0.01813 5.061 0.0562 0.1273 fixed
poly(age, 3, raw = TRUE)2 -0.002869 0.0006217 -4.615 -0.004088 -0.001651 fixed
poly(age, 3, raw = TRUE)3 0.00002989 0.000006824 4.379 0.00001651 0.00004326 fixed
male -0.02568 0.008478 -3.029 -0.04229 -0.00906 fixed
sibling_count3 0.01227 0.0158 0.7764 -0.0187 0.04323 fixed
sibling_count4 0.0249 0.01611 1.546 -0.00667 0.05646 fixed
sibling_count5 0.02296 0.01731 1.327 -0.01097 0.05689 fixed
sibling_count5+ -0.002852 0.01697 -0.168 -0.03611 0.03041 fixed
sd_(Intercept).mother_pidlink 0.1179 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2352 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.844 0.1687 -5.003 -1.175 -0.5133 fixed
poly(age, 3, raw = TRUE)1 0.08976 0.01812 4.953 0.05424 0.1253 fixed
poly(age, 3, raw = TRUE)2 -0.0028 0.0006216 -4.503 -0.004018 -0.001581 fixed
poly(age, 3, raw = TRUE)3 0.00002914 0.000006826 4.268 0.00001576 0.00004251 fixed
male -0.02612 0.008475 -3.082 -0.04273 -0.009508 fixed
sibling_count3 0.01501 0.01605 0.9355 -0.01644 0.04646 fixed
sibling_count4 0.02489 0.01654 1.505 -0.007523 0.05731 fixed
sibling_count5 0.02825 0.018 1.569 -0.007033 0.06353 fixed
sibling_count5+ -0.003534 0.01741 -0.203 -0.03766 0.03059 fixed
birth_order_nonlinear2 -0.0175 0.01125 -1.555 -0.03955 0.004559 fixed
birth_order_nonlinear3 -0.02054 0.01315 -1.561 -0.04632 0.005246 fixed
birth_order_nonlinear4 -0.001737 0.01557 -0.1115 -0.03226 0.02879 fixed
birth_order_nonlinear5 -0.05317 0.01898 -2.802 -0.09037 -0.01598 fixed
birth_order_nonlinear5+ -0.02094 0.01784 -1.174 -0.0559 0.01402 fixed
sd_(Intercept).mother_pidlink 0.1181 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.235 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.8285 0.1689 -4.904 -1.16 -0.4974 fixed
poly(age, 3, raw = TRUE)1 0.0882 0.01815 4.86 0.05263 0.1238 fixed
poly(age, 3, raw = TRUE)2 -0.002742 0.0006227 -4.404 -0.003963 -0.001522 fixed
poly(age, 3, raw = TRUE)3 0.00002847 0.000006839 4.163 0.00001507 0.00004188 fixed
male -0.0269 0.008477 -3.173 -0.04351 -0.01028 fixed
count_birth_order2/2 -0.0224 0.0238 -0.9412 -0.06904 0.02424 fixed
count_birth_order1/3 0.02072 0.02048 1.012 -0.01943 0.06087 fixed
count_birth_order2/3 -0.01094 0.02226 -0.4914 -0.05457 0.03269 fixed
count_birth_order3/3 -0.01167 0.02445 -0.4772 -0.05958 0.03625 fixed
count_birth_order1/4 0.01157 0.02226 0.5198 -0.03206 0.0552 fixed
count_birth_order2/4 0.0485 0.0232 2.09 0.003027 0.09397 fixed
count_birth_order3/4 -0.008379 0.02533 -0.3308 -0.05803 0.04127 fixed
count_birth_order4/4 -0.0125 0.02703 -0.4624 -0.06548 0.04048 fixed
count_birth_order1/5 0.03036 0.02665 1.139 -0.02187 0.08259 fixed
count_birth_order2/5 -0.02963 0.02765 -1.071 -0.08383 0.02457 fixed
count_birth_order3/5 -0.007984 0.02813 -0.2838 -0.06312 0.04715 fixed
count_birth_order4/5 0.07203 0.02859 2.52 0.016 0.1281 fixed
count_birth_order5/5 -0.01908 0.0288 -0.6625 -0.07552 0.03736 fixed
count_birth_order1/5+ -0.007641 0.0241 -0.317 -0.05488 0.0396 fixed
count_birth_order2/5+ -0.03434 0.02542 -1.351 -0.08417 0.01549 fixed
count_birth_order3/5+ -0.002912 0.02449 -0.1189 -0.05091 0.04509 fixed
count_birth_order4/5+ -0.01136 0.02388 -0.4756 -0.05817 0.03545 fixed
count_birth_order5/5+ -0.0629 0.02524 -2.492 -0.1124 -0.01343 fixed
count_birth_order5+/5+ -0.02624 0.01925 -1.364 -0.06396 0.01148 fixed
sd_(Intercept).mother_pidlink 0.1185 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2345 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 598.1 666.9 -288 576.1 NA NA NA
12 597.2 672.3 -286.6 573.2 2.897 1 0.08874
16 597.7 697.8 -282.8 565.7 7.516 4 0.111
26 599.1 761.8 -273.5 547.1 18.6 10 0.04565

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.8904 0.1716 -5.188 -1.227 -0.554 fixed
poly(age, 3, raw = TRUE)1 0.09498 0.01845 5.148 0.05882 0.1311 fixed
poly(age, 3, raw = TRUE)2 -0.002992 0.0006334 -4.723 -0.004233 -0.00175 fixed
poly(age, 3, raw = TRUE)3 0.00003151 0.000006961 4.527 0.00001787 0.00004515 fixed
male -0.02612 0.008613 -3.032 -0.043 -0.009236 fixed
sibling_count3 0.005929 0.01414 0.4193 -0.02178 0.03364 fixed
sibling_count4 0.01116 0.01487 0.7506 -0.01799 0.04031 fixed
sibling_count5 0.009306 0.01708 0.5449 -0.02417 0.04278 fixed
sibling_count5+ -0.03426 0.01459 -2.347 -0.06286 -0.005654 fixed
sd_(Intercept).mother_pidlink 0.1183 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.236 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.8946 0.1717 -5.211 -1.231 -0.5582 fixed
birth_order -0.003066 0.002754 -1.113 -0.008464 0.002332 fixed
poly(age, 3, raw = TRUE)1 0.09584 0.01847 5.19 0.05965 0.132 fixed
poly(age, 3, raw = TRUE)2 -0.003015 0.0006337 -4.757 -0.004257 -0.001773 fixed
poly(age, 3, raw = TRUE)3 0.00003162 0.000006962 4.543 0.00001798 0.00004527 fixed
male -0.02608 0.008613 -3.028 -0.04297 -0.009201 fixed
sibling_count3 0.007558 0.01421 0.5318 -0.0203 0.03541 fixed
sibling_count4 0.01457 0.01518 0.9599 -0.01518 0.04432 fixed
sibling_count5 0.01484 0.01778 0.8345 -0.02001 0.04969 fixed
sibling_count5+ -0.02312 0.0177 -1.306 -0.0578 0.01157 fixed
sd_(Intercept).mother_pidlink 0.118 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2361 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.8848 0.1719 -5.148 -1.222 -0.548 fixed
poly(age, 3, raw = TRUE)1 0.0948 0.01847 5.132 0.05859 0.131 fixed
poly(age, 3, raw = TRUE)2 -0.002977 0.0006341 -4.695 -0.00422 -0.001734 fixed
poly(age, 3, raw = TRUE)3 0.00003121 0.000006968 4.48 0.00001756 0.00004487 fixed
male -0.0264 0.008617 -3.063 -0.04329 -0.009507 fixed
sibling_count3 0.009706 0.01449 0.6698 -0.01869 0.0381 fixed
sibling_count4 0.01476 0.01569 0.9406 -0.01599 0.0455 fixed
sibling_count5 0.01424 0.0185 0.7696 -0.02202 0.0505 fixed
sibling_count5+ -0.02438 0.01816 -1.343 -0.05996 0.01121 fixed
birth_order_nonlinear2 -0.01562 0.0111 -1.408 -0.03737 0.006128 fixed
birth_order_nonlinear3 -0.01657 0.01315 -1.26 -0.04234 0.009199 fixed
birth_order_nonlinear4 -0.002065 0.0165 -0.1252 -0.0344 0.03027 fixed
birth_order_nonlinear5 -0.0181 0.02032 -0.891 -0.05792 0.02172 fixed
birth_order_nonlinear5+ -0.02136 0.0204 -1.047 -0.06135 0.01863 fixed
sd_(Intercept).mother_pidlink 0.1181 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2362 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.8625 0.1725 -5 -1.201 -0.5244 fixed
poly(age, 3, raw = TRUE)1 0.09284 0.01855 5.006 0.05649 0.1292 fixed
poly(age, 3, raw = TRUE)2 -0.002912 0.0006367 -4.574 -0.00416 -0.001665 fixed
poly(age, 3, raw = TRUE)3 0.00003055 0.000006999 4.365 0.00001684 0.00004427 fixed
male -0.02659 0.008634 -3.08 -0.04351 -0.00967 fixed
count_birth_order2/2 -0.02698 0.02109 -1.28 -0.06831 0.01435 fixed
count_birth_order1/3 0.01649 0.01833 0.8995 -0.01944 0.05241 fixed
count_birth_order2/3 -0.01704 0.02034 -0.8376 -0.05691 0.02283 fixed
count_birth_order3/3 -0.02125 0.02176 -0.9768 -0.06389 0.02139 fixed
count_birth_order1/4 -0.01368 0.02147 -0.6371 -0.05575 0.0284 fixed
count_birth_order2/4 0.0192 0.0224 0.857 -0.02471 0.0631 fixed
count_birth_order3/4 0.004533 0.02346 0.1932 -0.04145 0.05051 fixed
count_birth_order4/4 0.003269 0.02529 0.1293 -0.0463 0.05284 fixed
count_birth_order1/5 0.00178 0.02805 0.06347 -0.05319 0.05675 fixed
count_birth_order2/5 -0.01452 0.03139 -0.4625 -0.07603 0.047 fixed
count_birth_order3/5 -0.009643 0.03007 -0.3207 -0.06858 0.0493 fixed
count_birth_order4/5 0.02184 0.02901 0.753 -0.03502 0.0787 fixed
count_birth_order5/5 0.0006849 0.03053 0.02243 -0.05915 0.06052 fixed
count_birth_order1/5+ -0.02163 0.0273 -0.7922 -0.07513 0.03188 fixed
count_birth_order2/5+ -0.04792 0.0276 -1.736 -0.102 0.006187 fixed
count_birth_order3/5+ -0.03897 0.02643 -1.474 -0.09076 0.01283 fixed
count_birth_order4/5+ -0.03379 0.02627 -1.286 -0.08528 0.0177 fixed
count_birth_order5/5+ -0.05096 0.02487 -2.049 -0.09971 -0.002216 fixed
count_birth_order5+/5+ -0.0495 0.01961 -2.524 -0.08793 -0.01106 fixed
sd_(Intercept).mother_pidlink 0.1183 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2362 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 608.1 676.7 -293 586.1 NA NA NA
12 608.8 683.6 -292.4 584.8 1.244 1 0.2646
16 614.6 714.3 -291.3 582.6 2.283 4 0.6838
26 627.7 789.7 -287.8 575.7 6.885 10 0.7363

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Category_Private worker

birthorder <- birthorder %>% mutate(outcome = `Category_Private worker`)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.02918 0.1054 -0.2769 -0.2357 0.1774 fixed
poly(age, 3, raw = TRUE)1 0.05562 0.00946 5.88 0.03708 0.07416 fixed
poly(age, 3, raw = TRUE)2 -0.001725 0.000266 -6.484 -0.002246 -0.001203 fixed
poly(age, 3, raw = TRUE)3 0.00001406 0.000002366 5.941 0.00000942 0.0000187 fixed
male 0.05526 0.00963 5.738 0.03639 0.07414 fixed
sibling_count3 0.0003855 0.02066 0.01865 -0.04012 0.04089 fixed
sibling_count4 -0.008148 0.02089 -0.39 -0.04909 0.03279 fixed
sibling_count5 -0.02683 0.02182 -1.229 -0.0696 0.01595 fixed
sibling_count5+ -0.0556 0.01685 -3.3 -0.08863 -0.02258 fixed
sd_(Intercept).mother_pidlink 0.2157 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4361 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.02842 0.1054 -0.2697 -0.235 0.1781 fixed
birth_order -0.002304 0.00199 -1.158 -0.006205 0.001597 fixed
poly(age, 3, raw = TRUE)1 0.05627 0.009476 5.938 0.0377 0.07485 fixed
poly(age, 3, raw = TRUE)2 -0.00175 0.0002668 -6.556 -0.002273 -0.001227 fixed
poly(age, 3, raw = TRUE)3 0.00001429 0.000002374 6.017 0.000009633 0.00001894 fixed
male 0.05529 0.00963 5.741 0.03641 0.07416 fixed
sibling_count3 0.0009427 0.02067 0.04561 -0.03957 0.04145 fixed
sibling_count4 -0.006841 0.02092 -0.327 -0.04784 0.03416 fixed
sibling_count5 -0.02445 0.02192 -1.116 -0.06741 0.0185 fixed
sibling_count5+ -0.04769 0.01819 -2.622 -0.08333 -0.01204 fixed
sd_(Intercept).mother_pidlink 0.2156 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4362 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.02558 0.1055 -0.2424 -0.2324 0.1813 fixed
poly(age, 3, raw = TRUE)1 0.0563 0.009484 5.937 0.03771 0.07489 fixed
poly(age, 3, raw = TRUE)2 -0.001739 0.000267 -6.513 -0.002262 -0.001215 fixed
poly(age, 3, raw = TRUE)3 0.00001409 0.000002375 5.932 0.000009435 0.00001875 fixed
male 0.0553 0.009632 5.741 0.03642 0.07417 fixed
sibling_count3 0.003277 0.02094 0.1565 -0.03777 0.04432 fixed
sibling_count4 -0.003538 0.02142 -0.1652 -0.04552 0.03844 fixed
sibling_count5 -0.02212 0.02258 -0.9795 -0.06639 0.02214 fixed
sibling_count5+ -0.04829 0.01897 -2.546 -0.08548 -0.01111 fixed
birth_order_nonlinear2 -0.02808 0.01393 -2.016 -0.05538 -0.000778 fixed
birth_order_nonlinear3 -0.02067 0.01617 -1.278 -0.05235 0.01102 fixed
birth_order_nonlinear4 -0.02172 0.01816 -1.196 -0.05731 0.01388 fixed
birth_order_nonlinear5 -0.01489 0.02044 -0.7284 -0.05494 0.02517 fixed
birth_order_nonlinear5+ -0.02235 0.01718 -1.301 -0.05603 0.01132 fixed
sd_(Intercept).mother_pidlink 0.2154 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4363 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.02217 0.1059 -0.2094 -0.2298 0.1854 fixed
poly(age, 3, raw = TRUE)1 0.05677 0.009494 5.98 0.03816 0.07538 fixed
poly(age, 3, raw = TRUE)2 -0.001744 0.0002672 -6.528 -0.002268 -0.001221 fixed
poly(age, 3, raw = TRUE)3 0.00001406 0.000002377 5.916 0.000009403 0.00001872 fixed
male 0.05551 0.009637 5.76 0.03662 0.0744 fixed
count_birth_order2/2 -0.05724 0.02767 -2.068 -0.1115 -0.003003 fixed
count_birth_order1/3 0.0003303 0.02726 0.01212 -0.05309 0.05375 fixed
count_birth_order2/3 -0.03185 0.03017 -1.056 -0.09097 0.02727 fixed
count_birth_order3/3 -0.04859 0.03309 -1.469 -0.1134 0.01626 fixed
count_birth_order1/4 -0.005195 0.02981 -0.1743 -0.06363 0.05324 fixed
count_birth_order2/4 -0.03899 0.03194 -1.221 -0.1016 0.02361 fixed
count_birth_order3/4 -0.0355 0.03402 -1.043 -0.1022 0.03118 fixed
count_birth_order4/4 -0.0587 0.03611 -1.626 -0.1295 0.01207 fixed
count_birth_order1/5 -0.02741 0.0338 -0.8109 -0.09366 0.03884 fixed
count_birth_order2/5 -0.06241 0.03582 -1.742 -0.1326 0.007795 fixed
count_birth_order3/5 -0.07078 0.03768 -1.879 -0.1446 0.003066 fixed
count_birth_order4/5 -0.03619 0.03954 -0.9152 -0.1137 0.04131 fixed
count_birth_order5/5 -0.0552 0.03957 -1.395 -0.1327 0.02235 fixed
count_birth_order1/5+ -0.08694 0.026 -3.343 -0.1379 -0.03597 fixed
count_birth_order2/5+ -0.07812 0.02706 -2.887 -0.1312 -0.02509 fixed
count_birth_order3/5+ -0.06474 0.02665 -2.429 -0.117 -0.0125 fixed
count_birth_order4/5+ -0.07889 0.0262 -3.012 -0.1302 -0.02755 fixed
count_birth_order5/5+ -0.07251 0.02647 -2.739 -0.1244 -0.02062 fixed
count_birth_order5+/5+ -0.08181 0.0217 -3.77 -0.1243 -0.03927 fixed
sd_(Intercept).mother_pidlink 0.2159 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4362 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 13876 13956 -6927 13854 NA NA NA
12 13877 13964 -6927 13853 1.343 1 0.2466
16 13882 13997 -6925 13850 3.301 4 0.5088
26 13895 14083 -6922 13843 6.494 10 0.7722

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.026 0.3156 -6.419 -2.644 -1.407 fixed
poly(age, 3, raw = TRUE)1 0.2738 0.03387 8.083 0.2074 0.3402 fixed
poly(age, 3, raw = TRUE)2 -0.008982 0.001162 -7.732 -0.01126 -0.006705 fixed
poly(age, 3, raw = TRUE)3 0.00009119 0.00001275 7.151 0.0000662 0.0001162 fixed
male 0.03804 0.01584 2.401 0.00698 0.06909 fixed
sibling_count3 -0.03885 0.0269 -1.444 -0.09158 0.01387 fixed
sibling_count4 -0.03699 0.02791 -1.325 -0.09169 0.01771 fixed
sibling_count5 -0.07355 0.03119 -2.358 -0.1347 -0.01243 fixed
sibling_count5+ -0.1243 0.02709 -4.59 -0.1774 -0.07124 fixed
sd_(Intercept).mother_pidlink 0.2278 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4351 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.023 0.3157 -6.408 -2.642 -1.404 fixed
birth_order 0.002475 0.00498 0.4971 -0.007285 0.01224 fixed
poly(age, 3, raw = TRUE)1 0.2732 0.0339 8.058 0.2067 0.3396 fixed
poly(age, 3, raw = TRUE)2 -0.008966 0.001162 -7.714 -0.01124 -0.006688 fixed
poly(age, 3, raw = TRUE)3 0.00009112 0.00001275 7.145 0.00006613 0.0001161 fixed
male 0.03797 0.01585 2.396 0.006908 0.06903 fixed
sibling_count3 -0.04015 0.02703 -1.485 -0.09314 0.01283 fixed
sibling_count4 -0.0398 0.02848 -1.398 -0.09562 0.01602 fixed
sibling_count5 -0.07815 0.03254 -2.402 -0.1419 -0.01438 fixed
sibling_count5+ -0.1334 0.03264 -4.087 -0.1973 -0.06941 fixed
sd_(Intercept).mother_pidlink 0.228 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4351 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.027 0.3161 -6.411 -2.646 -1.407 fixed
poly(age, 3, raw = TRUE)1 0.2741 0.03392 8.08 0.2076 0.3405 fixed
poly(age, 3, raw = TRUE)2 -0.009001 0.001163 -7.739 -0.01128 -0.006721 fixed
poly(age, 3, raw = TRUE)3 0.00009161 0.00001277 7.176 0.00006659 0.0001166 fixed
male 0.0381 0.01585 2.403 0.007026 0.06917 fixed
sibling_count3 -0.03972 0.02752 -1.443 -0.09365 0.01421 fixed
sibling_count4 -0.03822 0.0294 -1.3 -0.09584 0.0194 fixed
sibling_count5 -0.08413 0.03393 -2.48 -0.1506 -0.01764 fixed
sibling_count5+ -0.1433 0.03342 -4.287 -0.2088 -0.07778 fixed
birth_order_nonlinear2 -0.00469 0.02067 -0.2269 -0.0452 0.03582 fixed
birth_order_nonlinear3 0.002097 0.02435 0.08612 -0.04563 0.04982 fixed
birth_order_nonlinear4 -0.001416 0.0297 -0.04766 -0.05964 0.0568 fixed
birth_order_nonlinear5 0.04611 0.03654 1.262 -0.0255 0.1177 fixed
birth_order_nonlinear5+ 0.0272 0.03676 0.7398 -0.04486 0.09925 fixed
sd_(Intercept).mother_pidlink 0.2275 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4354 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.078 0.3173 -6.548 -2.7 -1.456 fixed
poly(age, 3, raw = TRUE)1 0.2796 0.03406 8.211 0.2129 0.3464 fixed
poly(age, 3, raw = TRUE)2 -0.009203 0.001168 -7.879 -0.01149 -0.006913 fixed
poly(age, 3, raw = TRUE)3 0.00009392 0.00001282 7.324 0.00006879 0.0001191 fixed
male 0.03892 0.01588 2.451 0.007802 0.07005 fixed
count_birth_order2/2 0.005079 0.04026 0.1262 -0.07383 0.08399 fixed
count_birth_order1/3 -0.02901 0.03467 -0.8367 -0.09695 0.03894 fixed
count_birth_order2/3 -0.04598 0.03843 -1.197 -0.1213 0.02933 fixed
count_birth_order3/3 -0.0427 0.04151 -1.029 -0.1241 0.03866 fixed
count_birth_order1/4 -0.01582 0.03976 -0.3978 -0.09375 0.06211 fixed
count_birth_order2/4 -0.03676 0.04174 -0.8806 -0.1186 0.04506 fixed
count_birth_order3/4 -0.02532 0.04372 -0.579 -0.111 0.06038 fixed
count_birth_order4/4 -0.08068 0.04636 -1.74 -0.1715 0.01018 fixed
count_birth_order1/5 -0.07601 0.05206 -1.46 -0.178 0.02603 fixed
count_birth_order2/5 -0.1138 0.05651 -2.014 -0.2246 -0.00306 fixed
count_birth_order3/5 -0.03707 0.05297 -0.6999 -0.1409 0.06674 fixed
count_birth_order4/5 -0.1023 0.05152 -1.986 -0.2033 -0.001362 fixed
count_birth_order5/5 -0.03525 0.054 -0.6528 -0.1411 0.07058 fixed
count_birth_order1/5+ -0.1897 0.04946 -3.835 -0.2866 -0.09274 fixed
count_birth_order2/5+ -0.13 0.04983 -2.609 -0.2277 -0.03232 fixed
count_birth_order3/5+ -0.1686 0.0487 -3.462 -0.264 -0.07314 fixed
count_birth_order4/5+ -0.08096 0.04728 -1.712 -0.1736 0.0117 fixed
count_birth_order5/5+ -0.09381 0.04561 -2.057 -0.1832 -0.004414 fixed
count_birth_order5+/5+ -0.1121 0.03598 -3.115 -0.1826 -0.04156 fixed
sd_(Intercept).mother_pidlink 0.2282 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4353 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 5366 5435 -2672 5344 NA NA NA
12 5368 5443 -2672 5344 0.2466 1 0.6195
16 5374 5474 -2671 5342 2.082 4 0.7207
26 5386 5549 -2667 5334 8.045 10 0.6245

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.018 0.3145 -6.417 -2.635 -1.402 fixed
poly(age, 3, raw = TRUE)1 0.272 0.03379 8.048 0.2057 0.3382 fixed
poly(age, 3, raw = TRUE)2 -0.008924 0.001159 -7.699 -0.0112 -0.006652 fixed
poly(age, 3, raw = TRUE)3 0.00009054 0.00001273 7.114 0.00006559 0.0001155 fixed
male 0.03847 0.0158 2.434 0.007495 0.06945 fixed
sibling_count3 -0.01225 0.02948 -0.4156 -0.07003 0.04553 fixed
sibling_count4 -0.02102 0.0298 -0.7053 -0.07942 0.03739 fixed
sibling_count5 -0.04856 0.03152 -1.541 -0.1103 0.01322 fixed
sibling_count5+ -0.09925 0.02759 -3.598 -0.1533 -0.04518 fixed
sd_(Intercept).mother_pidlink 0.2287 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4351 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.02 0.3146 -6.42 -2.636 -1.403 fixed
birth_order -0.001348 0.004409 -0.3058 -0.009989 0.007293 fixed
poly(age, 3, raw = TRUE)1 0.2723 0.03381 8.053 0.206 0.3386 fixed
poly(age, 3, raw = TRUE)2 -0.008932 0.00116 -7.703 -0.0112 -0.00666 fixed
poly(age, 3, raw = TRUE)3 0.00009057 0.00001273 7.115 0.00006562 0.0001155 fixed
male 0.0385 0.01581 2.436 0.007519 0.06948 fixed
sibling_count3 -0.01153 0.02958 -0.3896 -0.0695 0.04645 fixed
sibling_count4 -0.01958 0.03017 -0.6491 -0.07871 0.03954 fixed
sibling_count5 -0.04622 0.03244 -1.425 -0.1098 0.01735 fixed
sibling_count5+ -0.09442 0.03179 -2.971 -0.1567 -0.03212 fixed
sd_(Intercept).mother_pidlink 0.2286 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4352 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.011 0.315 -6.386 -2.629 -1.394 fixed
poly(age, 3, raw = TRUE)1 0.2716 0.03383 8.03 0.2053 0.3379 fixed
poly(age, 3, raw = TRUE)2 -0.008913 0.00116 -7.682 -0.01119 -0.006639 fixed
poly(age, 3, raw = TRUE)3 0.00009045 0.00001274 7.1 0.00006548 0.0001154 fixed
male 0.03842 0.01581 2.43 0.007433 0.06941 fixed
sibling_count3 -0.01008 0.03005 -0.3354 -0.06897 0.04881 fixed
sibling_count4 -0.01969 0.03098 -0.6358 -0.08041 0.04102 fixed
sibling_count5 -0.05592 0.03372 -1.658 -0.122 0.01017 fixed
sibling_count5+ -0.103 0.03261 -3.158 -0.1669 -0.03907 fixed
birth_order_nonlinear2 -0.01341 0.02097 -0.6398 -0.05451 0.02768 fixed
birth_order_nonlinear3 -0.01003 0.02451 -0.409 -0.05807 0.03802 fixed
birth_order_nonlinear4 0.000195 0.02903 0.006718 -0.0567 0.05709 fixed
birth_order_nonlinear5 0.04181 0.03537 1.182 -0.02752 0.1111 fixed
birth_order_nonlinear5+ -0.009151 0.0333 -0.2748 -0.07443 0.05612 fixed
sd_(Intercept).mother_pidlink 0.228 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4355 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.009 0.3158 -6.363 -2.628 -1.39 fixed
poly(age, 3, raw = TRUE)1 0.2733 0.03392 8.058 0.2068 0.3398 fixed
poly(age, 3, raw = TRUE)2 -0.008987 0.001164 -7.723 -0.01127 -0.006706 fixed
poly(age, 3, raw = TRUE)3 0.00009141 0.00001278 7.153 0.00006637 0.0001165 fixed
male 0.03913 0.01584 2.471 0.008089 0.07016 fixed
count_birth_order2/2 -0.05345 0.04441 -1.204 -0.1405 0.03359 fixed
count_birth_order1/3 -0.01953 0.03832 -0.5096 -0.09464 0.05558 fixed
count_birth_order2/3 -0.04084 0.04163 -0.981 -0.1224 0.04075 fixed
count_birth_order3/3 -0.03316 0.04571 -0.7255 -0.1228 0.05643 fixed
count_birth_order1/4 -0.04812 0.04163 -1.156 -0.1297 0.03348 fixed
count_birth_order2/4 -0.03541 0.04338 -0.8163 -0.1204 0.04962 fixed
count_birth_order3/4 -0.01073 0.04735 -0.2265 -0.1035 0.08208 fixed
count_birth_order4/4 -0.05686 0.05053 -1.125 -0.1559 0.04216 fixed
count_birth_order1/5 -0.04768 0.04983 -0.9569 -0.1453 0.04998 fixed
count_birth_order2/5 -0.08619 0.05169 -1.667 -0.1875 0.01512 fixed
count_birth_order3/5 -0.05817 0.05257 -1.106 -0.1612 0.04488 fixed
count_birth_order4/5 -0.129 0.05342 -2.416 -0.2338 -0.02435 fixed
count_birth_order5/5 -0.009836 0.05381 -0.1828 -0.1153 0.09563 fixed
count_birth_order1/5+ -0.1407 0.04506 -3.123 -0.229 -0.0524 fixed
count_birth_order2/5+ -0.1041 0.04751 -2.191 -0.1972 -0.01097 fixed
count_birth_order3/5+ -0.1657 0.04577 -3.621 -0.2554 -0.07601 fixed
count_birth_order4/5+ -0.06336 0.04463 -1.42 -0.1508 0.02411 fixed
count_birth_order5/5+ -0.08578 0.04716 -1.819 -0.1782 0.00664 fixed
count_birth_order5+/5+ -0.1241 0.03602 -3.446 -0.1947 -0.05354 fixed
sd_(Intercept).mother_pidlink 0.2294 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.435 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 5409 5477 -2693 5387 NA NA NA
12 5410 5486 -2693 5386 0.09448 1 0.7586
16 5416 5516 -2692 5384 2.809 4 0.5902
26 5426 5589 -2687 5374 9.447 10 0.4903

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.997 0.3192 -6.256 -2.622 -1.371 fixed
poly(age, 3, raw = TRUE)1 0.2714 0.03431 7.912 0.2042 0.3387 fixed
poly(age, 3, raw = TRUE)2 -0.008929 0.001178 -7.582 -0.01124 -0.006621 fixed
poly(age, 3, raw = TRUE)3 0.00009099 0.00001294 7.031 0.00006562 0.0001164 fixed
male 0.0349 0.01601 2.18 0.003529 0.06628 fixed
sibling_count3 -0.04555 0.02637 -1.727 -0.09725 0.006137 fixed
sibling_count4 -0.04056 0.02775 -1.462 -0.09495 0.01383 fixed
sibling_count5 -0.0787 0.03189 -2.468 -0.1412 -0.01619 fixed
sibling_count5+ -0.1247 0.02725 -4.577 -0.1781 -0.07132 fixed
sd_(Intercept).mother_pidlink 0.227 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.436 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.994 0.3193 -6.245 -2.62 -1.368 fixed
birth_order 0.002076 0.005125 0.405 -0.00797 0.01212 fixed
poly(age, 3, raw = TRUE)1 0.2708 0.03434 7.887 0.2035 0.3381 fixed
poly(age, 3, raw = TRUE)2 -0.008913 0.001178 -7.564 -0.01122 -0.006603 fixed
poly(age, 3, raw = TRUE)3 0.00009091 0.00001294 7.023 0.00006554 0.0001163 fixed
male 0.03489 0.01601 2.18 0.003515 0.06627 fixed
sibling_count3 -0.04667 0.02652 -1.76 -0.09865 0.005312 fixed
sibling_count4 -0.04288 0.02834 -1.513 -0.09842 0.01266 fixed
sibling_count5 -0.08244 0.03322 -2.482 -0.1475 -0.01734 fixed
sibling_count5+ -0.1323 0.03304 -4.004 -0.1971 -0.06753 fixed
sd_(Intercept).mother_pidlink 0.2272 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.436 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.999 0.3196 -6.253 -2.625 -1.372 fixed
poly(age, 3, raw = TRUE)1 0.2719 0.03435 7.916 0.2046 0.3393 fixed
poly(age, 3, raw = TRUE)2 -0.008956 0.001179 -7.596 -0.01127 -0.006645 fixed
poly(age, 3, raw = TRUE)3 0.0000915 0.00001295 7.063 0.00006611 0.0001169 fixed
male 0.03526 0.01602 2.201 0.003867 0.06665 fixed
sibling_count3 -0.04768 0.02702 -1.764 -0.1006 0.005292 fixed
sibling_count4 -0.04006 0.02926 -1.369 -0.09741 0.0173 fixed
sibling_count5 -0.08599 0.03453 -2.49 -0.1537 -0.01831 fixed
sibling_count5+ -0.1441 0.03388 -4.252 -0.2105 -0.07766 fixed
birth_order_nonlinear2 -0.008968 0.02059 -0.4354 -0.04933 0.0314 fixed
birth_order_nonlinear3 0.006384 0.02441 0.2616 -0.04145 0.05422 fixed
birth_order_nonlinear4 -0.01746 0.03063 -0.5702 -0.0775 0.04257 fixed
birth_order_nonlinear5 0.03934 0.03771 1.043 -0.03458 0.1133 fixed
birth_order_nonlinear5+ 0.03292 0.03794 0.8678 -0.04143 0.1073 fixed
sd_(Intercept).mother_pidlink 0.2269 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4362 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -2.036 0.3208 -6.347 -2.665 -1.407 fixed
poly(age, 3, raw = TRUE)1 0.2765 0.03449 8.018 0.2089 0.3441 fixed
poly(age, 3, raw = TRUE)2 -0.009123 0.001184 -7.705 -0.01144 -0.006803 fixed
poly(age, 3, raw = TRUE)3 0.00009344 0.00001301 7.18 0.00006793 0.0001189 fixed
male 0.03622 0.01605 2.257 0.004772 0.06768 fixed
count_birth_order2/2 -0.01423 0.03915 -0.3636 -0.09096 0.0625 fixed
count_birth_order1/3 -0.04654 0.03411 -1.364 -0.1134 0.02032 fixed
count_birth_order2/3 -0.06353 0.03785 -1.678 -0.1377 0.01065 fixed
count_birth_order3/3 -0.04164 0.04047 -1.029 -0.121 0.03768 fixed
count_birth_order1/4 -0.02181 0.03994 -0.546 -0.1001 0.05648 fixed
count_birth_order2/4 -0.03984 0.04167 -0.9562 -0.1215 0.04183 fixed
count_birth_order3/4 -0.03702 0.04363 -0.8485 -0.1225 0.04849 fixed
count_birth_order4/4 -0.1064 0.04703 -2.262 -0.1986 -0.0142 fixed
count_birth_order1/5 -0.08905 0.05217 -1.707 -0.1913 0.0132 fixed
count_birth_order2/5 -0.1309 0.05835 -2.243 -0.2452 -0.0165 fixed
count_birth_order3/5 -0.04651 0.05591 -0.8318 -0.1561 0.06307 fixed
count_birth_order4/5 -0.1053 0.05394 -1.953 -0.2111 0.0003803 fixed
count_birth_order5/5 -0.04859 0.05676 -0.856 -0.1598 0.06266 fixed
count_birth_order1/5+ -0.1966 0.05077 -3.873 -0.2961 -0.09713 fixed
count_birth_order2/5+ -0.1305 0.05132 -2.542 -0.2311 -0.02987 fixed
count_birth_order3/5+ -0.1642 0.04914 -3.341 -0.2605 -0.06785 fixed
count_birth_order4/5+ -0.1132 0.04884 -2.318 -0.2089 -0.0175 fixed
count_birth_order5/5+ -0.1064 0.04624 -2.302 -0.1971 -0.01581 fixed
count_birth_order5+/5+ -0.112 0.03652 -3.067 -0.1836 -0.04044 fixed
sd_(Intercept).mother_pidlink 0.2272 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4363 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 5279 5348 -2629 5257 NA NA NA
12 5281 5356 -2628 5257 0.1631 1 0.6863
16 5286 5386 -2627 5254 2.91 4 0.573
26 5299 5461 -2624 5247 6.613 10 0.7614

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Category_Self-employed

birthorder <- birthorder %>% mutate(outcome = `Category_Self-employed`)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.2665 0.095 -2.805 -0.4527 -0.08029 fixed
poly(age, 3, raw = TRUE)1 0.02362 0.008521 2.772 0.006922 0.04033 fixed
poly(age, 3, raw = TRUE)2 -0.0001646 0.0002393 -0.6879 -0.0006336 0.0003044 fixed
poly(age, 3, raw = TRUE)3 -0.0000002153 0.000002126 -0.1013 -0.000004382 0.000003952 fixed
male -0.002977 0.008766 -0.3396 -0.02016 0.0142 fixed
sibling_count3 -0.02246 0.01828 -1.229 -0.05828 0.01337 fixed
sibling_count4 -0.01987 0.01844 -1.077 -0.056 0.01627 fixed
sibling_count5 -0.01023 0.01921 -0.5326 -0.04789 0.02743 fixed
sibling_count5+ 0.01732 0.0149 1.162 -0.01188 0.04651 fixed
sd_(Intercept).mother_pidlink 0.1531 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4098 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.2668 0.09499 -2.809 -0.453 -0.08063 fixed
birth_order 0.002029 0.001784 1.137 -0.001469 0.005526 fixed
poly(age, 3, raw = TRUE)1 0.02307 0.008535 2.703 0.006339 0.0398 fixed
poly(age, 3, raw = TRUE)2 -0.0001441 0.00024 -0.6006 -0.0006145 0.0003262 fixed
poly(age, 3, raw = TRUE)3 -0.0000004013 0.000002132 -0.1882 -0.00000458 0.000003778 fixed
male -0.003007 0.008766 -0.3431 -0.02019 0.01417 fixed
sibling_count3 -0.023 0.01828 -1.258 -0.05884 0.01283 fixed
sibling_count4 -0.0211 0.01847 -1.143 -0.0573 0.01509 fixed
sibling_count5 -0.01241 0.01931 -0.6426 -0.05025 0.02544 fixed
sibling_count5+ 0.01022 0.01615 0.6328 -0.02144 0.04188 fixed
sd_(Intercept).mother_pidlink 0.1529 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4098 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.2752 0.09511 -2.893 -0.4616 -0.08877 fixed
poly(age, 3, raw = TRUE)1 0.02319 0.008541 2.715 0.00645 0.03993 fixed
poly(age, 3, raw = TRUE)2 -0.0001598 0.0002401 -0.6657 -0.0006303 0.0003107 fixed
poly(age, 3, raw = TRUE)3 -0.0000001582 0.000002133 -0.07415 -0.000004338 0.000004022 fixed
male -0.003111 0.008765 -0.3549 -0.02029 0.01407 fixed
sibling_count3 -0.02462 0.01855 -1.327 -0.06098 0.01173 fixed
sibling_count4 -0.02199 0.01896 -1.16 -0.05915 0.01517 fixed
sibling_count5 -0.01296 0.01996 -0.6493 -0.05208 0.02616 fixed
sibling_count5+ 0.01309 0.01691 0.7741 -0.02005 0.04623 fixed
birth_order_nonlinear2 0.03823 0.01276 2.997 0.01323 0.06324 fixed
birth_order_nonlinear3 0.01961 0.01483 1.322 -0.009457 0.04867 fixed
birth_order_nonlinear4 0.0146 0.01665 0.877 -0.01803 0.04724 fixed
birth_order_nonlinear5 0.01968 0.01874 1.05 -0.01704 0.0564 fixed
birth_order_nonlinear5+ 0.02218 0.01558 1.423 -0.008359 0.05271 fixed
sd_(Intercept).mother_pidlink 0.1529 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4097 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.2834 0.09546 -2.969 -0.4705 -0.09631 fixed
poly(age, 3, raw = TRUE)1 0.02291 0.008548 2.68 0.006152 0.03966 fixed
poly(age, 3, raw = TRUE)2 -0.0001582 0.0002402 -0.6585 -0.000629 0.0003126 fixed
poly(age, 3, raw = TRUE)3 -0.000000111 0.000002134 -0.052 -0.000004293 0.000004072 fixed
male -0.003372 0.008769 -0.3845 -0.02056 0.01382 fixed
count_birth_order2/2 0.07421 0.02534 2.928 0.02454 0.1239 fixed
count_birth_order1/3 -0.02418 0.02456 -0.9844 -0.07231 0.02396 fixed
count_birth_order2/3 0.03323 0.02722 1.221 -0.02011 0.08657 fixed
count_birth_order3/3 0.02459 0.02989 0.8229 -0.03398 0.08317 fixed
count_birth_order1/4 0.01501 0.0269 0.5579 -0.03772 0.06773 fixed
count_birth_order2/4 0.005201 0.02884 0.1804 -0.05132 0.06173 fixed
count_birth_order3/4 -0.005191 0.03075 -0.1688 -0.06545 0.05507 fixed
count_birth_order4/4 0.01914 0.03266 0.5862 -0.04486 0.08315 fixed
count_birth_order1/5 0.009117 0.03054 0.2985 -0.05074 0.06897 fixed
count_birth_order2/5 0.04643 0.03238 1.434 -0.01704 0.1099 fixed
count_birth_order3/5 0.03461 0.03409 1.015 -0.0322 0.1014 fixed
count_birth_order4/5 -0.01429 0.03583 -0.3989 -0.08451 0.05593 fixed
count_birth_order5/5 0.0102 0.03582 0.2846 -0.06001 0.0804 fixed
count_birth_order1/5+ 0.03492 0.02352 1.485 -0.01118 0.08102 fixed
count_birth_order2/5+ 0.05428 0.02449 2.216 0.006268 0.1023 fixed
count_birth_order3/5+ 0.04115 0.02413 1.705 -0.006149 0.08844 fixed
count_birth_order4/5+ 0.04521 0.0237 1.907 -0.001244 0.09166 fixed
count_birth_order5/5+ 0.04958 0.02396 2.069 0.002613 0.09655 fixed
count_birth_order5+/5+ 0.04898 0.01941 2.523 0.01094 0.08703 fixed
sd_(Intercept).mother_pidlink 0.153 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4097 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 11871 11950 -5924 11849 NA NA NA
12 11871 11958 -5924 11847 1.295 1 0.2552
16 11872 11987 -5920 11840 7.919 4 0.0946
26 11883 12071 -5916 11831 8.217 10 0.6077

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.9428 0.2612 3.609 0.4307 1.455 fixed
poly(age, 3, raw = TRUE)1 -0.1058 0.02805 -3.773 -0.1608 -0.05087 fixed
poly(age, 3, raw = TRUE)2 0.004147 0.0009623 4.309 0.002261 0.006033 fixed
poly(age, 3, raw = TRUE)3 -0.00004614 0.00001056 -4.368 -0.00006685 -0.00002544 fixed
male -0.02367 0.01315 -1.8 -0.04943 0.0021 fixed
sibling_count3 0.01552 0.02194 0.7075 -0.02747 0.05851 fixed
sibling_count4 -0.02907 0.02271 -1.28 -0.07358 0.01545 fixed
sibling_count5 0.03856 0.02531 1.524 -0.01104 0.08815 fixed
sibling_count5+ 0.04773 0.02198 2.172 0.00465 0.09081 fixed
sd_(Intercept).mother_pidlink 0.1578 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3721 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.9413 0.2613 3.602 0.4291 1.453 fixed
birth_order -0.001447 0.004107 -0.3524 -0.009497 0.006603 fixed
poly(age, 3, raw = TRUE)1 -0.1055 0.02807 -3.759 -0.1605 -0.05049 fixed
poly(age, 3, raw = TRUE)2 0.004138 0.0009626 4.299 0.002251 0.006025 fixed
poly(age, 3, raw = TRUE)3 -0.00004611 0.00001057 -4.364 -0.00006682 -0.0000254 fixed
male -0.02362 0.01315 -1.796 -0.04939 0.002152 fixed
sibling_count3 0.01627 0.02204 0.7382 -0.02693 0.05948 fixed
sibling_count4 -0.02744 0.02318 -1.184 -0.07288 0.01799 fixed
sibling_count5 0.0412 0.02642 1.559 -0.01058 0.09299 fixed
sibling_count5+ 0.05295 0.02654 1.995 0.0009406 0.105 fixed
sd_(Intercept).mother_pidlink 0.1582 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.372 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.9329 0.2617 3.565 0.42 1.446 fixed
poly(age, 3, raw = TRUE)1 -0.1051 0.02809 -3.742 -0.1602 -0.05006 fixed
poly(age, 3, raw = TRUE)2 0.004127 0.0009634 4.283 0.002238 0.006015 fixed
poly(age, 3, raw = TRUE)3 -0.00004607 0.00001058 -4.355 -0.0000668 -0.00002534 fixed
male -0.02352 0.01315 -1.788 -0.0493 0.002267 fixed
sibling_count3 0.0156 0.02247 0.6942 -0.02844 0.05964 fixed
sibling_count4 -0.02767 0.02398 -1.154 -0.07467 0.01933 fixed
sibling_count5 0.04284 0.02763 1.55 -0.01132 0.09699 fixed
sibling_count5+ 0.0613 0.02722 2.252 0.007956 0.1147 fixed
birth_order_nonlinear2 0.01025 0.01727 0.5934 -0.02361 0.04411 fixed
birth_order_nonlinear3 0.001374 0.02033 0.06759 -0.03848 0.04123 fixed
birth_order_nonlinear4 -0.002639 0.02479 -0.1065 -0.05122 0.04594 fixed
birth_order_nonlinear5 -0.008891 0.03048 -0.2917 -0.06864 0.05086 fixed
birth_order_nonlinear5+ -0.02557 0.03043 -0.8404 -0.08522 0.03407 fixed
sd_(Intercept).mother_pidlink 0.1581 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3722 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.9316 0.2628 3.545 0.4166 1.447 fixed
poly(age, 3, raw = TRUE)1 -0.1065 0.02821 -3.774 -0.1618 -0.05117 fixed
poly(age, 3, raw = TRUE)2 0.004174 0.0009678 4.313 0.002277 0.006071 fixed
poly(age, 3, raw = TRUE)3 -0.00004658 0.00001063 -4.383 -0.00006741 -0.00002575 fixed
male -0.02415 0.01318 -1.832 -0.04998 0.001684 fixed
count_birth_order2/2 0.05231 0.0336 1.557 -0.01355 0.1182 fixed
count_birth_order1/3 0.03981 0.02861 1.392 -0.01626 0.09589 fixed
count_birth_order2/3 0.03881 0.03176 1.222 -0.02344 0.1011 fixed
count_birth_order3/3 0.01045 0.03434 0.3044 -0.05685 0.07775 fixed
count_birth_order1/4 -0.008259 0.03285 -0.2514 -0.07264 0.05612 fixed
count_birth_order2/4 -0.02985 0.03453 -0.8646 -0.09752 0.03782 fixed
count_birth_order3/4 -0.01272 0.0362 -0.3513 -0.08366 0.05823 fixed
count_birth_order4/4 0.008705 0.03839 0.2267 -0.06654 0.08395 fixed
count_birth_order1/5 0.05321 0.04307 1.235 -0.03121 0.1376 fixed
count_birth_order2/5 0.07654 0.04684 1.634 -0.01527 0.1684 fixed
count_birth_order3/5 0.06749 0.04389 1.538 -0.01854 0.1535 fixed
count_birth_order4/5 0.05473 0.04269 1.282 -0.02895 0.1384 fixed
count_birth_order5/5 0.02978 0.04476 0.6652 -0.05796 0.1175 fixed
count_birth_order1/5+ 0.07785 0.04095 1.901 -0.002414 0.1581 fixed
count_birth_order2/5+ 0.07465 0.04131 1.807 -0.006317 0.1556 fixed
count_birth_order3/5+ 0.09799 0.04038 2.426 0.01884 0.1771 fixed
count_birth_order4/5+ 0.04506 0.03922 1.149 -0.0318 0.1219 fixed
count_birth_order5/5+ 0.07709 0.03782 2.038 0.00296 0.1512 fixed
count_birth_order5+/5+ 0.04905 0.0296 1.657 -0.008953 0.1071 fixed
sd_(Intercept).mother_pidlink 0.1579 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3725 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 3908 3977 -1943 3886 NA NA NA
12 3910 3985 -1943 3886 0.122 1 0.7269
16 3917 4017 -1942 3885 1.352 4 0.8525
26 3931 4094 -1940 3879 5.638 10 0.8447

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.917 0.2603 3.523 0.4068 1.427 fixed
poly(age, 3, raw = TRUE)1 -0.1019 0.02798 -3.642 -0.1567 -0.04707 fixed
poly(age, 3, raw = TRUE)2 0.004027 0.0009598 4.196 0.002146 0.005909 fixed
poly(age, 3, raw = TRUE)3 -0.00004496 0.00001054 -4.266 -0.00006562 -0.0000243 fixed
male -0.02421 0.01311 -1.847 -0.04991 0.001481 fixed
sibling_count3 -0.006695 0.02405 -0.2784 -0.05383 0.04044 fixed
sibling_count4 -0.03614 0.02428 -1.489 -0.08372 0.01144 fixed
sibling_count5 -0.008562 0.02563 -0.3341 -0.05879 0.04167 fixed
sibling_count5+ 0.0278 0.02245 1.238 -0.0162 0.0718 fixed
sd_(Intercept).mother_pidlink 0.1587 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.372 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.9174 0.2603 3.524 0.4071 1.428 fixed
birth_order 0.0004435 0.003628 0.1222 -0.006668 0.007555 fixed
poly(age, 3, raw = TRUE)1 -0.102 0.02799 -3.644 -0.1568 -0.04713 fixed
poly(age, 3, raw = TRUE)2 0.00403 0.0009601 4.197 0.002148 0.005911 fixed
poly(age, 3, raw = TRUE)3 -0.00004497 0.00001054 -4.266 -0.00006563 -0.00002431 fixed
male -0.02422 0.01311 -1.848 -0.04992 0.001474 fixed
sibling_count3 -0.006931 0.02413 -0.2872 -0.05422 0.04036 fixed
sibling_count4 -0.03661 0.02457 -1.49 -0.08477 0.01156 fixed
sibling_count5 -0.009323 0.02638 -0.3534 -0.06102 0.04238 fixed
sibling_count5+ 0.02623 0.02589 1.013 -0.02451 0.07696 fixed
sd_(Intercept).mother_pidlink 0.1587 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3721 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.9044 0.2607 3.47 0.3935 1.415 fixed
poly(age, 3, raw = TRUE)1 -0.1011 0.02801 -3.608 -0.156 -0.04616 fixed
poly(age, 3, raw = TRUE)2 0.003996 0.0009609 4.159 0.002113 0.00588 fixed
poly(age, 3, raw = TRUE)3 -0.00004462 0.00001055 -4.229 -0.00006531 -0.00002394 fixed
male -0.02403 0.01312 -1.832 -0.04974 0.001677 fixed
sibling_count3 -0.01015 0.02454 -0.4134 -0.05824 0.03795 fixed
sibling_count4 -0.03844 0.02528 -1.521 -0.08799 0.0111 fixed
sibling_count5 -0.008874 0.02749 -0.3228 -0.06275 0.04501 fixed
sibling_count5+ 0.02904 0.02661 1.091 -0.02311 0.08119 fixed
birth_order_nonlinear2 0.01753 0.01751 1.001 -0.01679 0.05185 fixed
birth_order_nonlinear3 0.01558 0.02045 0.7616 -0.02451 0.05566 fixed
birth_order_nonlinear4 -0.00005852 0.0242 -0.002418 -0.04749 0.04737 fixed
birth_order_nonlinear5 -0.002062 0.02949 -0.06994 -0.05986 0.05573 fixed
birth_order_nonlinear5+ 0.002764 0.02752 0.1004 -0.05118 0.05671 fixed
sd_(Intercept).mother_pidlink 0.1585 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3722 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.904 0.2612 3.46 0.392 1.416 fixed
poly(age, 3, raw = TRUE)1 -0.1033 0.02807 -3.68 -0.1583 -0.04828 fixed
poly(age, 3, raw = TRUE)2 0.004077 0.0009634 4.233 0.002189 0.005966 fixed
poly(age, 3, raw = TRUE)3 -0.00004555 0.00001058 -4.305 -0.00006629 -0.00002481 fixed
male -0.02482 0.01313 -1.89 -0.05057 0.0009199 fixed
count_birth_order2/2 0.07947 0.03699 2.148 0.006965 0.152 fixed
count_birth_order1/3 0.02032 0.0316 0.643 -0.04161 0.08225 fixed
count_birth_order2/3 0.04303 0.03437 1.252 -0.02433 0.1104 fixed
count_birth_order3/3 -0.015 0.03778 -0.3971 -0.08904 0.05904 fixed
count_birth_order1/4 0.005795 0.03435 0.1687 -0.06153 0.07312 fixed
count_birth_order2/4 -0.0307 0.03584 -0.8566 -0.1009 0.03954 fixed
count_birth_order3/4 -0.0108 0.03917 -0.2758 -0.08757 0.06596 fixed
count_birth_order4/4 -0.006404 0.04181 -0.1532 -0.08835 0.07554 fixed
count_birth_order1/5 0.01609 0.04117 0.3908 -0.0646 0.09678 fixed
count_birth_order2/5 0.01692 0.04276 0.3957 -0.06688 0.1007 fixed
count_birth_order3/5 0.01622 0.04352 0.3728 -0.06907 0.1015 fixed
count_birth_order4/5 0.01776 0.04424 0.4013 -0.06896 0.1045 fixed
count_birth_order5/5 0.02105 0.04457 0.4723 -0.0663 0.1084 fixed
count_birth_order1/5+ 0.03557 0.03723 0.9554 -0.0374 0.1085 fixed
count_birth_order2/5+ 0.04765 0.03933 1.211 -0.02944 0.1247 fixed
count_birth_order3/5+ 0.1171 0.03789 3.092 0.04288 0.1914 fixed
count_birth_order4/5+ 0.03739 0.03695 1.012 -0.03504 0.1098 fixed
count_birth_order5/5+ 0.03913 0.03908 1.001 -0.03746 0.1157 fixed
count_birth_order5+/5+ 0.05174 0.02961 1.747 -0.006299 0.1098 fixed
sd_(Intercept).mother_pidlink 0.1582 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3723 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 3939 4008 -1958 3917 NA NA NA
12 3941 4016 -1958 3917 0.01549 1 0.901
16 3947 4047 -1958 3915 1.543 4 0.8189
26 3955 4118 -1952 3903 11.8 10 0.2988

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.9712 0.2646 3.67 0.4525 1.49 fixed
poly(age, 3, raw = TRUE)1 -0.11 0.02846 -3.864 -0.1657 -0.05418 fixed
poly(age, 3, raw = TRUE)2 0.004334 0.0009771 4.435 0.002418 0.006249 fixed
poly(age, 3, raw = TRUE)3 -0.00004871 0.00001074 -4.536 -0.00006976 -0.00002766 fixed
male -0.02276 0.0133 -1.711 -0.04883 0.003314 fixed
sibling_count3 0.02252 0.02157 1.044 -0.01976 0.06479 fixed
sibling_count4 -0.02657 0.02265 -1.173 -0.07097 0.01783 fixed
sibling_count5 0.03133 0.02594 1.208 -0.01952 0.08218 fixed
sibling_count5+ 0.04105 0.02217 1.852 -0.002396 0.0845 fixed
sd_(Intercept).mother_pidlink 0.16 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3726 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.9698 0.2647 3.663 0.4509 1.489 fixed
birth_order -0.001067 0.004238 -0.2517 -0.009373 0.007239 fixed
poly(age, 3, raw = TRUE)1 -0.1097 0.02848 -3.851 -0.1655 -0.05385 fixed
poly(age, 3, raw = TRUE)2 0.004326 0.0009776 4.425 0.00241 0.006242 fixed
poly(age, 3, raw = TRUE)3 -0.00004867 0.00001074 -4.531 -0.00006973 -0.00002762 fixed
male -0.02275 0.0133 -1.71 -0.04882 0.00333 fixed
sibling_count3 0.02308 0.02169 1.064 -0.01943 0.0656 fixed
sibling_count4 -0.02539 0.02314 -1.097 -0.07074 0.01996 fixed
sibling_count5 0.03322 0.02704 1.229 -0.01977 0.08622 fixed
sibling_count5+ 0.0449 0.02695 1.666 -0.007923 0.09772 fixed
sd_(Intercept).mother_pidlink 0.1602 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3725 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.9616 0.2651 3.628 0.4421 1.481 fixed
poly(age, 3, raw = TRUE)1 -0.1091 0.0285 -3.828 -0.165 -0.05323 fixed
poly(age, 3, raw = TRUE)2 0.00431 0.0009784 4.405 0.002392 0.006227 fixed
poly(age, 3, raw = TRUE)3 -0.00004859 0.00001075 -4.519 -0.00006966 -0.00002751 fixed
male -0.02281 0.01331 -1.713 -0.0489 0.003284 fixed
sibling_count3 0.02313 0.02213 1.045 -0.02025 0.06651 fixed
sibling_count4 -0.02585 0.02394 -1.08 -0.07278 0.02107 fixed
sibling_count5 0.03196 0.02818 1.134 -0.02328 0.08719 fixed
sibling_count5+ 0.05248 0.02768 1.896 -0.001773 0.1067 fixed
birth_order_nonlinear2 0.004606 0.01724 0.2672 -0.02918 0.03839 fixed
birth_order_nonlinear3 -0.001029 0.02041 -0.0504 -0.04103 0.03897 fixed
birth_order_nonlinear4 0.001722 0.02559 0.06727 -0.04844 0.05189 fixed
birth_order_nonlinear5 0.003465 0.03152 0.1099 -0.05832 0.06525 fixed
birth_order_nonlinear5+ -0.02943 0.03147 -0.9352 -0.09112 0.03225 fixed
sd_(Intercept).mother_pidlink 0.1602 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3727 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.9503 0.266 3.573 0.429 1.472 fixed
poly(age, 3, raw = TRUE)1 -0.1092 0.0286 -3.819 -0.1653 -0.05318 fixed
poly(age, 3, raw = TRUE)2 0.004316 0.0009823 4.394 0.002391 0.006241 fixed
poly(age, 3, raw = TRUE)3 -0.00004867 0.0000108 -4.507 -0.00006983 -0.00002751 fixed
male -0.0234 0.01334 -1.754 -0.04954 0.002743 fixed
count_birth_order2/2 0.0426 0.03271 1.302 -0.02152 0.1067 fixed
count_birth_order1/3 0.04535 0.0282 1.608 -0.009916 0.1006 fixed
count_birth_order2/3 0.04406 0.03134 1.406 -0.01736 0.1055 fixed
count_birth_order3/3 0.01164 0.03353 0.3472 -0.05408 0.07737 fixed
count_birth_order1/4 -0.008358 0.03305 -0.2528 -0.07314 0.05643 fixed
count_birth_order2/4 -0.04096 0.03452 -1.187 -0.1086 0.0267 fixed
count_birth_order3/4 -0.01007 0.03618 -0.2784 -0.08097 0.06083 fixed
count_birth_order4/4 0.02124 0.03901 0.5443 -0.05522 0.09769 fixed
count_birth_order1/5 0.0574 0.04322 1.328 -0.02731 0.1421 fixed
count_birth_order2/5 0.07243 0.04845 1.495 -0.02254 0.1674 fixed
count_birth_order3/5 0.0487 0.04642 1.049 -0.04227 0.1397 fixed
count_birth_order4/5 0.03821 0.04477 0.8534 -0.04954 0.126 fixed
count_birth_order5/5 0.0122 0.04712 0.2589 -0.08016 0.1046 fixed
count_birth_order1/5+ 0.05652 0.0421 1.342 -0.026 0.139 fixed
count_birth_order2/5+ 0.05222 0.04261 1.225 -0.0313 0.1357 fixed
count_birth_order3/5+ 0.09224 0.04079 2.261 0.01228 0.1722 fixed
count_birth_order4/5+ 0.03775 0.04057 0.9304 -0.04177 0.1173 fixed
count_birth_order5/5+ 0.0893 0.0384 2.326 0.01404 0.1646 fixed
count_birth_order5+/5+ 0.0354 0.0301 1.176 -0.02359 0.09438 fixed
sd_(Intercept).mother_pidlink 0.1598 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3729 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 3861 3930 -1919 3839 NA NA NA
12 3863 3938 -1919 3839 0.06169 1 0.8038
16 3870 3969 -1919 3838 1.332 4 0.856
26 3882 4044 -1915 3830 8.08 10 0.6211

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Category_Unpaid family worker

birthorder <- birthorder %>% mutate(outcome = `Category_Unpaid family worker`)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.174 0.06645 17.66 1.044 1.304 fixed
poly(age, 3, raw = TRUE)1 -0.0784 0.005963 -13.15 -0.09008 -0.06671 fixed
poly(age, 3, raw = TRUE)2 0.001808 0.0001676 10.79 0.001479 0.002136 fixed
poly(age, 3, raw = TRUE)3 -0.00001311 0.00000149 -8.799 -0.00001603 -0.00001019 fixed
male -0.09204 0.006103 -15.08 -0.104 -0.08008 fixed
sibling_count3 0.01356 0.0129 1.051 -0.01173 0.03885 fixed
sibling_count4 0.01864 0.01303 1.43 -0.006901 0.04417 fixed
sibling_count5 0.02536 0.0136 1.865 -0.001285 0.05201 fixed
sibling_count5+ 0.04657 0.01052 4.427 0.02595 0.06718 fixed
sd_(Intercept).mother_pidlink 0.1222 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2808 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.173 0.06645 17.66 1.043 1.303 fixed
birth_order 0.002099 0.001252 1.676 -0.000355 0.004553 fixed
poly(age, 3, raw = TRUE)1 -0.07898 0.005973 -13.22 -0.09069 -0.06727 fixed
poly(age, 3, raw = TRUE)2 0.00183 0.0001681 10.89 0.0015 0.002159 fixed
poly(age, 3, raw = TRUE)3 -0.00001331 0.000001494 -8.905 -0.00001624 -0.00001038 fixed
male -0.09207 0.006102 -15.09 -0.104 -0.08011 fixed
sibling_count3 0.01303 0.01291 1.009 -0.01227 0.03832 fixed
sibling_count4 0.0174 0.01305 1.333 -0.00818 0.04298 fixed
sibling_count5 0.02316 0.01366 1.695 -0.003617 0.04993 fixed
sibling_count5+ 0.03929 0.01138 3.453 0.01699 0.06159 fixed
sd_(Intercept).mother_pidlink 0.1223 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2807 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.178 0.06653 17.71 1.048 1.309 fixed
poly(age, 3, raw = TRUE)1 -0.07952 0.005977 -13.3 -0.09123 -0.0678 fixed
poly(age, 3, raw = TRUE)2 0.001839 0.0001681 10.94 0.00151 0.002169 fixed
poly(age, 3, raw = TRUE)3 -0.00001332 0.000001495 -8.914 -0.00001625 -0.00001039 fixed
male -0.09203 0.006102 -15.08 -0.104 -0.08007 fixed
sibling_count3 0.008595 0.01308 0.657 -0.01704 0.03423 fixed
sibling_count4 0.01064 0.01338 0.7952 -0.01558 0.03685 fixed
sibling_count5 0.01522 0.01409 1.08 -0.0124 0.04285 fixed
sibling_count5+ 0.0316 0.01189 2.658 0.008295 0.0549 fixed
birth_order_nonlinear2 0.01122 0.008852 1.268 -0.006126 0.02858 fixed
birth_order_nonlinear3 0.02609 0.01028 2.537 0.005937 0.04624 fixed
birth_order_nonlinear4 0.02517 0.01155 2.179 0.002532 0.0478 fixed
birth_order_nonlinear5 0.02458 0.01299 1.892 -0.0008849 0.05005 fixed
birth_order_nonlinear5+ 0.0255 0.01086 2.347 0.004204 0.04679 fixed
sd_(Intercept).mother_pidlink 0.122 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2808 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 1.183 0.06675 17.73 1.053 1.314 fixed
poly(age, 3, raw = TRUE)1 -0.08008 0.005981 -13.39 -0.09181 -0.06836 fixed
poly(age, 3, raw = TRUE)2 0.00185 0.0001682 11 0.001521 0.00218 fixed
poly(age, 3, raw = TRUE)3 -0.00001336 0.000001495 -8.938 -0.0000163 -0.00001043 fixed
male -0.09221 0.006105 -15.1 -0.1042 -0.08024 fixed
count_birth_order2/2 0.01891 0.01758 1.076 -0.01555 0.05337 fixed
count_birth_order1/3 0.001989 0.01717 0.1158 -0.03166 0.03564 fixed
count_birth_order2/3 0.01413 0.01902 0.7428 -0.02315 0.0514 fixed
count_birth_order3/3 0.06725 0.02087 3.222 0.02634 0.1082 fixed
count_birth_order1/4 0.01351 0.01879 0.7191 -0.02332 0.05035 fixed
count_birth_order2/4 0.0194 0.02014 0.9631 -0.02008 0.05888 fixed
count_birth_order3/4 0.02846 0.02147 1.326 -0.01361 0.07054 fixed
count_birth_order4/4 0.06085 0.0228 2.67 0.01618 0.1055 fixed
count_birth_order1/5 0.0184 0.02132 0.8626 -0.0234 0.06019 fixed
count_birth_order2/5 0.04596 0.02261 2.033 0.001651 0.09026 fixed
count_birth_order3/5 0.02026 0.02379 0.8517 -0.02636 0.06689 fixed
count_birth_order4/5 0.0333 0.02499 1.333 -0.01567 0.08228 fixed
count_birth_order5/5 0.059 0.02499 2.361 0.01001 0.108 fixed
count_birth_order1/5+ 0.04611 0.01642 2.809 0.01393 0.07828 fixed
count_birth_order2/5+ 0.04523 0.01709 2.647 0.01174 0.07873 fixed
count_birth_order3/5+ 0.0602 0.01683 3.576 0.0272 0.09319 fixed
count_birth_order4/5+ 0.05521 0.01654 3.338 0.02279 0.08762 fixed
count_birth_order5/5+ 0.05467 0.01672 3.27 0.0219 0.08743 fixed
count_birth_order5+/5+ 0.06034 0.01362 4.429 0.03364 0.08704 fixed
sd_(Intercept).mother_pidlink 0.1218 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2808 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 4606 4686 -2292 4584 NA NA NA
12 4605 4692 -2291 4581 2.811 1 0.09361
16 4607 4722 -2287 4575 6.69 4 0.1532
26 4616 4804 -2282 4564 10.22 10 0.4212

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 2.442 0.188 12.99 2.074 2.81 fixed
poly(age, 3, raw = TRUE)1 -0.2215 0.02019 -10.97 -0.2611 -0.1819 fixed
poly(age, 3, raw = TRUE)2 0.006628 0.0006926 9.57 0.00527 0.007985 fixed
poly(age, 3, raw = TRUE)3 -0.00006461 0.000007603 -8.497 -0.00007951 -0.00004971 fixed
male -0.03876 0.009466 -4.095 -0.05731 -0.02021 fixed
sibling_count3 0.02684 0.01573 1.706 -0.003987 0.05766 fixed
sibling_count4 0.05053 0.01628 3.105 0.01863 0.08243 fixed
sibling_count5 0.05469 0.01812 3.018 0.01918 0.0902 fixed
sibling_count5+ 0.07393 0.01574 4.699 0.04309 0.1048 fixed
sd_(Intercept).mother_pidlink 0.1072 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2701 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 2.445 0.188 13 2.076 2.813 fixed
birth_order 0.003024 0.002952 1.024 -0.002762 0.008809 fixed
poly(age, 3, raw = TRUE)1 -0.2222 0.0202 -11 -0.2617 -0.1826 fixed
poly(age, 3, raw = TRUE)2 0.006644 0.0006927 9.591 0.005286 0.008002 fixed
poly(age, 3, raw = TRUE)3 -0.00006465 0.000007603 -8.503 -0.00007956 -0.00004975 fixed
male -0.03886 0.009466 -4.105 -0.05741 -0.02031 fixed
sibling_count3 0.02528 0.0158 1.6 -0.005696 0.05626 fixed
sibling_count4 0.04716 0.01661 2.839 0.0146 0.07971 fixed
sibling_count5 0.04914 0.01892 2.598 0.01206 0.08622 fixed
sibling_count5+ 0.06302 0.01901 3.315 0.02577 0.1003 fixed
sd_(Intercept).mother_pidlink 0.1075 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.27 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 2.444 0.1882 12.99 2.075 2.813 fixed
poly(age, 3, raw = TRUE)1 -0.222 0.0202 -10.99 -0.2616 -0.1824 fixed
poly(age, 3, raw = TRUE)2 0.006638 0.000693 9.578 0.005279 0.007996 fixed
poly(age, 3, raw = TRUE)3 -0.00006456 0.000007608 -8.486 -0.00007948 -0.00004965 fixed
male -0.03863 0.009467 -4.08 -0.05718 -0.02007 fixed
sibling_count3 0.01909 0.0161 1.186 -0.01247 0.05065 fixed
sibling_count4 0.03957 0.01718 2.303 0.005902 0.07325 fixed
sibling_count5 0.0444 0.01979 2.244 0.005624 0.08318 fixed
sibling_count5+ 0.05921 0.01949 3.038 0.02101 0.09741 fixed
birth_order_nonlinear2 0.01273 0.01246 1.022 -0.01169 0.03714 fixed
birth_order_nonlinear3 0.03267 0.01466 2.229 0.003941 0.0614 fixed
birth_order_nonlinear4 0.02027 0.01786 1.135 -0.01474 0.05528 fixed
birth_order_nonlinear5 0.004018 0.02197 0.1828 -0.03905 0.04708 fixed
birth_order_nonlinear5+ 0.02633 0.02188 1.203 -0.01656 0.06922 fixed
sd_(Intercept).mother_pidlink 0.107 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2701 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 2.448 0.1889 12.96 2.078 2.819 fixed
poly(age, 3, raw = TRUE)1 -0.2223 0.02028 -10.96 -0.2621 -0.1826 fixed
poly(age, 3, raw = TRUE)2 0.006654 0.0006958 9.563 0.00529 0.008018 fixed
poly(age, 3, raw = TRUE)3 -0.00006481 0.00000764 -8.483 -0.00007979 -0.00004984 fixed
male -0.03842 0.00948 -4.053 -0.057 -0.01984 fixed
count_birth_order2/2 0.003413 0.02421 0.141 -0.04403 0.05086 fixed
count_birth_order1/3 0.0001465 0.02055 0.007127 -0.04013 0.04043 fixed
count_birth_order2/3 0.03307 0.02282 1.449 -0.01166 0.0778 fixed
count_birth_order3/3 0.07378 0.02468 2.99 0.02541 0.1222 fixed
count_birth_order1/4 0.05301 0.0236 2.246 0.006748 0.09927 fixed
count_birth_order2/4 0.03648 0.02481 1.47 -0.01216 0.08511 fixed
count_birth_order3/4 0.04578 0.02602 1.759 -0.005222 0.09678 fixed
count_birth_order4/4 0.07534 0.0276 2.73 0.02124 0.1294 fixed
count_birth_order1/5 0.03696 0.03096 1.194 -0.02371 0.09764 fixed
count_birth_order2/5 0.09427 0.03368 2.799 0.02825 0.1603 fixed
count_birth_order3/5 0.05057 0.03156 1.602 -0.01129 0.1124 fixed
count_birth_order4/5 0.0572 0.0307 1.863 -0.002967 0.1174 fixed
count_birth_order5/5 0.04417 0.03219 1.372 -0.01891 0.1073 fixed
count_birth_order1/5+ 0.06705 0.02944 2.278 0.009357 0.1248 fixed
count_birth_order2/5+ 0.06292 0.0297 2.118 0.004706 0.1211 fixed
count_birth_order3/5+ 0.101 0.02904 3.479 0.0441 0.1579 fixed
count_birth_order4/5+ 0.06068 0.0282 2.152 0.005407 0.116 fixed
count_birth_order5/5+ 0.06101 0.0272 2.243 0.007709 0.1143 fixed
count_birth_order5+/5+ 0.08233 0.02124 3.876 0.04071 0.124 fixed
sd_(Intercept).mother_pidlink 0.107 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2701 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 1386 1455 -682.2 1364 NA NA NA
12 1387 1462 -681.6 1363 1.048 1 0.3059
16 1390 1490 -679.2 1358 4.901 4 0.2976
26 1401 1563 -674.3 1349 9.699 10 0.4672

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 2.464 0.1878 13.13 2.096 2.832 fixed
poly(age, 3, raw = TRUE)1 -0.2235 0.02018 -11.07 -0.263 -0.1839 fixed
poly(age, 3, raw = TRUE)2 0.00669 0.0006925 9.661 0.005332 0.008047 fixed
poly(age, 3, raw = TRUE)3 -0.0000652 0.000007604 -8.574 -0.0000801 -0.00005029 fixed
male -0.03937 0.009461 -4.161 -0.05792 -0.02083 fixed
sibling_count3 0.02618 0.01729 1.514 -0.007704 0.06007 fixed
sibling_count4 0.0433 0.01744 2.482 0.009104 0.07749 fixed
sibling_count5 0.03684 0.01841 2.001 0.0007636 0.07291 fixed
sibling_count5+ 0.06644 0.01613 4.12 0.03483 0.09805 fixed
sd_(Intercept).mother_pidlink 0.1083 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2705 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 2.468 0.1877 13.14 2.1 2.835 fixed
birth_order 0.004248 0.002612 1.626 -0.0008725 0.009368 fixed
poly(age, 3, raw = TRUE)1 -0.2243 0.02018 -11.11 -0.2639 -0.1848 fixed
poly(age, 3, raw = TRUE)2 0.00671 0.0006924 9.691 0.005353 0.008067 fixed
poly(age, 3, raw = TRUE)3 -0.00006523 0.000007602 -8.581 -0.00008013 -0.00005033 fixed
male -0.03948 0.009459 -4.174 -0.05802 -0.02094 fixed
sibling_count3 0.02393 0.01734 1.38 -0.01007 0.05792 fixed
sibling_count4 0.03886 0.01766 2.201 0.00425 0.07346 fixed
sibling_count5 0.02956 0.01894 1.561 -0.007566 0.06669 fixed
sibling_count5+ 0.05138 0.0186 2.763 0.01494 0.08783 fixed
sd_(Intercept).mother_pidlink 0.1085 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2704 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 2.458 0.1878 13.09 2.09 2.827 fixed
poly(age, 3, raw = TRUE)1 -0.2232 0.02018 -11.06 -0.2627 -0.1836 fixed
poly(age, 3, raw = TRUE)2 0.006673 0.0006925 9.636 0.005316 0.00803 fixed
poly(age, 3, raw = TRUE)3 -0.00006486 0.000007605 -8.528 -0.00007976 -0.00004995 fixed
male -0.03913 0.009457 -4.137 -0.05766 -0.02059 fixed
sibling_count3 0.01657 0.01763 0.9395 -0.018 0.05113 fixed
sibling_count4 0.0298 0.01816 1.641 -0.005797 0.0654 fixed
sibling_count5 0.02269 0.01975 1.149 -0.01601 0.0614 fixed
sibling_count5+ 0.04598 0.01911 2.406 0.008519 0.08344 fixed
birth_order_nonlinear2 0.01053 0.01265 0.8323 -0.01426 0.03531 fixed
birth_order_nonlinear3 0.03994 0.01477 2.705 0.011 0.06888 fixed
birth_order_nonlinear4 0.02845 0.01747 1.629 -0.00579 0.06269 fixed
birth_order_nonlinear5 0.009549 0.02128 0.4487 -0.03217 0.05127 fixed
birth_order_nonlinear5+ 0.03222 0.01982 1.625 -0.006636 0.07107 fixed
sd_(Intercept).mother_pidlink 0.1088 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2702 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 2.444 0.1881 12.99 2.075 2.813 fixed
poly(age, 3, raw = TRUE)1 -0.2221 0.02022 -10.99 -0.2617 -0.1825 fixed
poly(age, 3, raw = TRUE)2 0.00664 0.0006938 9.571 0.005281 0.008 fixed
poly(age, 3, raw = TRUE)3 -0.00006454 0.000007622 -8.468 -0.00007948 -0.00004961 fixed
male -0.03836 0.009465 -4.052 -0.05691 -0.01981 fixed
count_birth_order2/2 0.02215 0.02669 0.83 -0.03016 0.07447 fixed
count_birth_order1/3 0.00529 0.02274 0.2326 -0.03928 0.04986 fixed
count_birth_order2/3 0.02414 0.02474 0.9755 -0.02436 0.07263 fixed
count_birth_order3/3 0.09777 0.0272 3.594 0.04445 0.1511 fixed
count_birth_order1/4 0.05529 0.02473 2.236 0.006829 0.1038 fixed
count_birth_order2/4 0.01373 0.0258 0.5321 -0.03684 0.0643 fixed
count_birth_order3/4 0.06335 0.02821 2.246 0.008066 0.1186 fixed
count_birth_order4/4 0.08211 0.03012 2.727 0.02309 0.1411 fixed
count_birth_order1/5 0.001284 0.02964 0.04333 -0.05681 0.05938 fixed
count_birth_order2/5 0.08761 0.03079 2.845 0.02726 0.148 fixed
count_birth_order3/5 0.05313 0.03135 1.695 -0.008308 0.1146 fixed
count_birth_order4/5 0.05656 0.03187 1.775 -0.005906 0.119 fixed
count_birth_order5/5 0.02284 0.03211 0.7115 -0.04008 0.08577 fixed
count_birth_order1/5+ 0.07247 0.0268 2.704 0.01993 0.125 fixed
count_birth_order2/5+ 0.06204 0.02833 2.19 0.006505 0.1176 fixed
count_birth_order3/5+ 0.07122 0.02729 2.61 0.01773 0.1247 fixed
count_birth_order4/5+ 0.06325 0.02662 2.376 0.01108 0.1154 fixed
count_birth_order5/5+ 0.06806 0.02816 2.417 0.01288 0.1232 fixed
count_birth_order5+/5+ 0.08184 0.02129 3.844 0.04011 0.1236 fixed
sd_(Intercept).mother_pidlink 0.1072 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2705 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 1418 1487 -697.9 1396 NA NA NA
12 1417 1492 -696.5 1393 2.648 1 0.1037
16 1419 1519 -693.4 1387 6.263 4 0.1804
26 1422 1585 -685 1370 16.78 10 0.07945

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 2.439 0.1895 12.87 2.068 2.81 fixed
poly(age, 3, raw = TRUE)1 -0.2205 0.02038 -10.82 -0.2605 -0.1806 fixed
poly(age, 3, raw = TRUE)2 0.006576 0.0006996 9.4 0.005205 0.007948 fixed
poly(age, 3, raw = TRUE)3 -0.00006393 0.00000769 -8.314 -0.00007901 -0.00004886 fixed
male -0.03783 0.009531 -3.97 -0.05651 -0.01915 fixed
sibling_count3 0.02853 0.01537 1.857 -0.001589 0.05865 fixed
sibling_count4 0.04803 0.01613 2.978 0.01642 0.07964 fixed
sibling_count5 0.06003 0.01844 3.255 0.02389 0.09618 fixed
sibling_count5+ 0.07782 0.01576 4.938 0.04693 0.1087 fixed
sd_(Intercept).mother_pidlink 0.1061 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2697 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 2.442 0.1895 12.89 2.071 2.814 fixed
birth_order 0.002716 0.003029 0.8965 -0.003222 0.008654 fixed
poly(age, 3, raw = TRUE)1 -0.2212 0.02039 -10.85 -0.2612 -0.1812 fixed
poly(age, 3, raw = TRUE)2 0.006594 0.0006999 9.421 0.005222 0.007966 fixed
poly(age, 3, raw = TRUE)3 -0.00006401 0.00000769 -8.323 -0.00007908 -0.00004893 fixed
male -0.03786 0.009531 -3.972 -0.05654 -0.01918 fixed
sibling_count3 0.0271 0.01545 1.754 -0.003187 0.05739 fixed
sibling_count4 0.04504 0.01647 2.735 0.01276 0.07732 fixed
sibling_count5 0.05519 0.01922 2.871 0.01752 0.09287 fixed
sibling_count5+ 0.06803 0.01917 3.548 0.03045 0.1056 fixed
sd_(Intercept).mother_pidlink 0.1063 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2697 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 2.443 0.1897 12.88 2.071 2.815 fixed
poly(age, 3, raw = TRUE)1 -0.2215 0.02039 -10.86 -0.2615 -0.1815 fixed
poly(age, 3, raw = TRUE)2 0.006599 0.0007002 9.425 0.005227 0.007972 fixed
poly(age, 3, raw = TRUE)3 -0.00006404 0.000007695 -8.322 -0.00007912 -0.00004895 fixed
male -0.03763 0.009533 -3.948 -0.05632 -0.01895 fixed
sibling_count3 0.02229 0.01576 1.415 -0.008594 0.05318 fixed
sibling_count4 0.03807 0.01704 2.234 0.004672 0.07147 fixed
sibling_count5 0.05204 0.02004 2.598 0.01277 0.09131 fixed
sibling_count5+ 0.06606 0.01968 3.356 0.02748 0.1046 fixed
birth_order_nonlinear2 0.01948 0.01238 1.573 -0.004787 0.04374 fixed
birth_order_nonlinear3 0.027 0.01465 1.843 -0.001718 0.05571 fixed
birth_order_nonlinear4 0.02471 0.01837 1.345 -0.01129 0.06071 fixed
birth_order_nonlinear5 0.001399 0.02263 0.06181 -0.04295 0.04575 fixed
birth_order_nonlinear5+ 0.02411 0.02252 1.071 -0.02002 0.06824 fixed
sd_(Intercept).mother_pidlink 0.1055 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2699 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 2.451 0.1904 12.87 2.078 2.824 fixed
poly(age, 3, raw = TRUE)1 -0.2225 0.02047 -10.87 -0.2626 -0.1824 fixed
poly(age, 3, raw = TRUE)2 0.006642 0.0007031 9.448 0.005264 0.00802 fixed
poly(age, 3, raw = TRUE)3 -0.00006459 0.000007729 -8.357 -0.00007974 -0.00004945 fixed
male -0.03784 0.009553 -3.961 -0.05656 -0.01912 fixed
count_birth_order2/2 0.01989 0.0235 0.8465 -0.02616 0.06594 fixed
count_birth_order1/3 0.008942 0.02016 0.4434 -0.03058 0.04846 fixed
count_birth_order2/3 0.04507 0.02242 2.01 0.001131 0.08902 fixed
count_birth_order3/3 0.07044 0.024 2.935 0.02341 0.1175 fixed
count_birth_order1/4 0.05241 0.02364 2.217 0.006072 0.09875 fixed
count_birth_order2/4 0.04846 0.0247 1.962 0.00003909 0.09688 fixed
count_birth_order3/4 0.04797 0.02589 1.852 -0.002785 0.09872 fixed
count_birth_order4/4 0.07353 0.02793 2.633 0.0188 0.1283 fixed
count_birth_order1/5 0.05752 0.03092 1.86 -0.00309 0.1181 fixed
count_birth_order2/5 0.07077 0.0347 2.04 0.002764 0.1388 fixed
count_birth_order3/5 0.05268 0.03324 1.585 -0.01247 0.1178 fixed
count_birth_order4/5 0.07725 0.03206 2.41 0.01442 0.1401 fixed
count_birth_order5/5 0.07412 0.03375 2.196 0.00798 0.1403 fixed
count_birth_order1/5+ 0.07883 0.03014 2.616 0.01977 0.1379 fixed
count_birth_order2/5+ 0.09488 0.03051 3.11 0.03508 0.1547 fixed
count_birth_order3/5+ 0.1004 0.02921 3.436 0.04311 0.1576 fixed
count_birth_order4/5+ 0.07875 0.02906 2.71 0.02179 0.1357 fixed
count_birth_order5/5+ 0.05546 0.0275 2.017 0.001564 0.1094 fixed
count_birth_order5+/5+ 0.09005 0.02148 4.192 0.04795 0.1322 fixed
sd_(Intercept).mother_pidlink 0.1046 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2704 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 1344 1413 -661 1322 NA NA NA
12 1345 1420 -660.6 1321 0.8029 1 0.3702
16 1348 1448 -658.2 1316 4.708 4 0.3186
26 1362 1524 -655.2 1310 6.094 10 0.8073

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Work Sector

Sector_Agriculture, forestry, fishing and hunting

birthorder <- birthorder %>% mutate(outcome = `Sector_Agriculture, forestry, fishing and hunting`)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.1306 0.0897 1.456 -0.04522 0.3064 fixed
poly(age, 3, raw = TRUE)1 -0.0008862 0.008036 -0.1103 -0.01664 0.01486 fixed
poly(age, 3, raw = TRUE)2 0.0001282 0.0002255 0.5685 -0.0003138 0.0005702 fixed
poly(age, 3, raw = TRUE)3 -0.000001381 0.000002002 -0.6899 -0.000005305 0.000002542 fixed
male 0.02977 0.008213 3.624 0.01367 0.04586 fixed
sibling_count3 0.00264 0.01767 0.1494 -0.032 0.03728 fixed
sibling_count4 0.0012 0.01789 0.06708 -0.03386 0.03626 fixed
sibling_count5 0.00142 0.01864 0.07615 -0.03512 0.03796 fixed
sibling_count5+ 0.01776 0.01442 1.232 -0.01051 0.04604 fixed
sd_(Intercept).mother_pidlink 0.1829 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3648 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.1302 0.0897 1.451 -0.04566 0.306 fixed
birth_order 0.001497 0.001692 0.8848 -0.001819 0.004812 fixed
poly(age, 3, raw = TRUE)1 -0.001315 0.008051 -0.1634 -0.01709 0.01446 fixed
poly(age, 3, raw = TRUE)2 0.0001445 0.0002262 0.6385 -0.000299 0.0005879 fixed
poly(age, 3, raw = TRUE)3 -0.00000153 0.000002009 -0.7617 -0.000005467 0.000002407 fixed
male 0.02974 0.008213 3.621 0.01364 0.04584 fixed
sibling_count3 0.0023 0.01768 0.1301 -0.03235 0.03695 fixed
sibling_count4 0.0003682 0.01791 0.02055 -0.03474 0.03547 fixed
sibling_count5 -0.00009096 0.01872 -0.004859 -0.03678 0.0366 fixed
sibling_count5+ 0.01263 0.01555 0.812 -0.01785 0.0431 fixed
sd_(Intercept).mother_pidlink 0.1827 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3649 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.1386 0.08983 1.543 -0.03742 0.3147 fixed
poly(age, 3, raw = TRUE)1 -0.001996 0.008057 -0.2477 -0.01779 0.0138 fixed
poly(age, 3, raw = TRUE)2 0.0001671 0.0002264 0.7384 -0.0002765 0.0006108 fixed
poly(age, 3, raw = TRUE)3 -0.000001733 0.00000201 -0.8621 -0.000005672 0.000002206 fixed
male 0.02977 0.008214 3.624 0.01367 0.04587 fixed
sibling_count3 0.002077 0.01791 0.116 -0.03302 0.03717 fixed
sibling_count4 -0.003051 0.01834 -0.1663 -0.039 0.0329 fixed
sibling_count5 -0.00412 0.01928 -0.2137 -0.04191 0.03367 fixed
sibling_count5+ 0.004782 0.01621 0.295 -0.02699 0.03656 fixed
birth_order_nonlinear2 -0.004716 0.01189 -0.3967 -0.02802 0.01858 fixed
birth_order_nonlinear3 0.003489 0.01378 0.2532 -0.02352 0.0305 fixed
birth_order_nonlinear4 0.02163 0.0155 1.396 -0.008745 0.05201 fixed
birth_order_nonlinear5 0.0108 0.01742 0.6202 -0.02333 0.04494 fixed
birth_order_nonlinear5+ 0.02094 0.01463 1.431 -0.007737 0.04962 fixed
sd_(Intercept).mother_pidlink 0.1826 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3649 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.1623 0.09007 1.802 -0.01423 0.3388 fixed
poly(age, 3, raw = TRUE)1 -0.00313 0.008058 -0.3884 -0.01892 0.01266 fixed
poly(age, 3, raw = TRUE)2 0.0001942 0.0002263 0.8581 -0.0002494 0.0006378 fixed
poly(age, 3, raw = TRUE)3 -0.000001913 0.000002009 -0.9523 -0.000005851 0.000002024 fixed
male 0.02973 0.00821 3.622 0.01364 0.04582 fixed
count_birth_order2/2 -0.03256 0.02366 -1.376 -0.07893 0.01381 fixed
count_birth_order1/3 -0.02768 0.02327 -1.19 -0.07329 0.01793 fixed
count_birth_order2/3 -0.02653 0.02583 -1.027 -0.07717 0.0241 fixed
count_birth_order3/3 0.05159 0.02831 1.822 -0.003905 0.1071 fixed
count_birth_order1/4 -0.02445 0.02556 -0.9567 -0.07454 0.02564 fixed
count_birth_order2/4 -0.04134 0.02739 -1.509 -0.09502 0.01234 fixed
count_birth_order3/4 0.03614 0.02885 1.253 -0.0204 0.09268 fixed
count_birth_order4/4 0.007561 0.03102 0.2438 -0.05323 0.06836 fixed
count_birth_order1/5 -0.01458 0.02876 -0.507 -0.07096 0.04179 fixed
count_birth_order2/5 -0.001386 0.03052 -0.04544 -0.06119 0.05842 fixed
count_birth_order3/5 -0.04137 0.03206 -1.29 -0.1042 0.02148 fixed
count_birth_order4/5 0.04476 0.03378 1.325 -0.02145 0.111 fixed
count_birth_order5/5 -0.02623 0.03388 -0.7743 -0.09264 0.04017 fixed
count_birth_order1/5+ 0.006506 0.02222 0.2929 -0.03704 0.05005 fixed
count_birth_order2/5+ 0.02085 0.02307 0.9039 -0.02436 0.06606 fixed
count_birth_order3/5+ -0.04021 0.02283 -1.761 -0.08495 0.004531 fixed
count_birth_order4/5+ 0.007129 0.02239 0.3185 -0.03675 0.051 fixed
count_birth_order5/5+ 0.01262 0.02261 0.5581 -0.0317 0.05694 fixed
count_birth_order5+/5+ 0.01603 0.01858 0.8625 -0.02039 0.05244 fixed
sd_(Intercept).mother_pidlink 0.1828 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3644 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 9969 10049 -4974 9947 NA NA NA
12 9971 10057 -4973 9947 0.7849 1 0.3756
16 9975 10090 -4972 9943 3.619 4 0.4601
26 9966 10153 -4957 9914 28.92 10 0.001283

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.2048 0.2492 0.8218 -0.2836 0.6932 fixed
poly(age, 3, raw = TRUE)1 -0.006416 0.02672 -0.2401 -0.05878 0.04595 fixed
poly(age, 3, raw = TRUE)2 0.0002029 0.0009152 0.2217 -0.001591 0.001997 fixed
poly(age, 3, raw = TRUE)3 -0.00000112 0.00001003 -0.1116 -0.00002078 0.00001854 fixed
male 0.01815 0.01255 1.446 -0.006448 0.04275 fixed
sibling_count3 -0.01727 0.02117 -0.8156 -0.05875 0.02422 fixed
sibling_count4 -0.006193 0.02193 -0.2825 -0.04917 0.03678 fixed
sibling_count5 0.005548 0.02456 0.2259 -0.04259 0.05369 fixed
sibling_count5+ 0.01615 0.02122 0.7613 -0.02543 0.05773 fixed
sd_(Intercept).mother_pidlink 0.1644 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3417 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.2093 0.2493 0.8396 -0.2793 0.6979 fixed
birth_order 0.003167 0.003924 0.8071 -0.004523 0.01086 fixed
poly(age, 3, raw = TRUE)1 -0.007299 0.02674 -0.2729 -0.05971 0.04512 fixed
poly(age, 3, raw = TRUE)2 0.0002263 0.0009157 0.2472 -0.001568 0.002021 fixed
poly(age, 3, raw = TRUE)3 -0.000001234 0.00001003 -0.123 -0.0000209 0.00001843 fixed
male 0.01803 0.01255 1.437 -0.00657 0.04263 fixed
sibling_count3 -0.01886 0.02126 -0.8871 -0.06053 0.02281 fixed
sibling_count4 -0.009734 0.02236 -0.4353 -0.05356 0.0341 fixed
sibling_count5 -0.00025 0.02559 -0.009769 -0.05041 0.04991 fixed
sibling_count5+ 0.004629 0.02557 0.181 -0.04549 0.05475 fixed
sd_(Intercept).mother_pidlink 0.1644 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3417 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.2063 0.2494 0.8274 -0.2824 0.6951 fixed
poly(age, 3, raw = TRUE)1 -0.007303 0.02673 -0.2732 -0.05969 0.04509 fixed
poly(age, 3, raw = TRUE)2 0.0002221 0.0009154 0.2426 -0.001572 0.002016 fixed
poly(age, 3, raw = TRUE)3 -0.000001135 0.00001003 -0.1131 -0.0000208 0.00001853 fixed
male 0.01846 0.01255 1.471 -0.006134 0.04306 fixed
sibling_count3 -0.02808 0.02162 -1.299 -0.07046 0.0143 fixed
sibling_count4 -0.02 0.02309 -0.8659 -0.06526 0.02526 fixed
sibling_count5 -0.004471 0.02665 -0.1678 -0.05671 0.04777 fixed
sibling_count5+ -0.002136 0.02617 -0.08163 -0.05342 0.04915 fixed
birth_order_nonlinear2 0.02484 0.01643 1.512 -0.00737 0.05705 fixed
birth_order_nonlinear3 0.04716 0.01931 2.443 0.009321 0.085 fixed
birth_order_nonlinear4 0.02073 0.02359 0.8784 -0.02552 0.06697 fixed
birth_order_nonlinear5 -0.005032 0.02894 -0.1739 -0.06176 0.05169 fixed
birth_order_nonlinear5+ 0.04153 0.02893 1.435 -0.01518 0.09824 fixed
sd_(Intercept).mother_pidlink 0.164 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3417 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.1737 0.2502 0.6943 -0.3167 0.6641 fixed
poly(age, 3, raw = TRUE)1 -0.003514 0.02683 -0.131 -0.0561 0.04907 fixed
poly(age, 3, raw = TRUE)2 0.00007322 0.000919 0.07967 -0.001728 0.001874 fixed
poly(age, 3, raw = TRUE)3 0.0000006975 0.00001008 0.06922 -0.00001905 0.00002045 fixed
male 0.01881 0.01256 1.498 -0.005805 0.04343 fixed
count_birth_order2/2 0.03164 0.03213 0.9849 -0.03133 0.09461 fixed
count_birth_order1/3 -0.007422 0.02734 -0.2715 -0.06101 0.04616 fixed
count_birth_order2/3 -0.0317 0.03035 -1.045 -0.09118 0.02777 fixed
count_birth_order3/3 0.0254 0.03301 0.7693 -0.0393 0.0901 fixed
count_birth_order1/4 -0.0155 0.03133 -0.4946 -0.0769 0.04591 fixed
count_birth_order2/4 0.009061 0.03316 0.2732 -0.05593 0.07406 fixed
count_birth_order3/4 0.05157 0.03435 1.501 -0.01576 0.1189 fixed
count_birth_order4/4 -0.02999 0.03675 -0.8159 -0.102 0.04205 fixed
count_birth_order1/5 -0.007561 0.0412 -0.1835 -0.08831 0.07319 fixed
count_birth_order2/5 -0.0012 0.04434 -0.02707 -0.0881 0.0857 fixed
count_birth_order3/5 0.04731 0.04215 1.123 -0.03529 0.1299 fixed
count_birth_order4/5 0.05736 0.0411 1.396 -0.0232 0.1379 fixed
count_birth_order5/5 -0.02381 0.04323 -0.5508 -0.1085 0.06092 fixed
count_birth_order1/5+ -0.0458 0.03899 -1.175 -0.1222 0.03062 fixed
count_birth_order2/5+ 0.09322 0.03922 2.377 0.01634 0.1701 fixed
count_birth_order3/5+ 0.01043 0.03826 0.2726 -0.06456 0.08543 fixed
count_birth_order4/5+ 0.02492 0.03743 0.6657 -0.04844 0.09828 fixed
count_birth_order5/5+ 0.005131 0.03579 0.1434 -0.06502 0.07528 fixed
count_birth_order5+/5+ 0.0424 0.02827 1.5 -0.01302 0.09782 fixed
sd_(Intercept).mother_pidlink 0.1621 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3422 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 3263 3332 -1621 3241 NA NA NA
12 3265 3339 -1620 3241 0.6535 1 0.4189
16 3265 3364 -1616 3233 7.902 4 0.09525
26 3271 3432 -1609 3219 14.38 10 0.1564

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.2 0.248 0.8065 -0.286 0.686 fixed
poly(age, 3, raw = TRUE)1 -0.005978 0.02661 -0.2246 -0.05814 0.04618 fixed
poly(age, 3, raw = TRUE)2 0.0001901 0.0009118 0.2085 -0.001597 0.001977 fixed
poly(age, 3, raw = TRUE)3 -0.0000009767 0.000009998 -0.09769 -0.00002057 0.00001862 fixed
male 0.01874 0.0125 1.499 -0.005762 0.04324 fixed
sibling_count3 -0.02278 0.02317 -0.9834 -0.06819 0.02262 fixed
sibling_count4 -0.004367 0.02342 -0.1865 -0.05027 0.04154 fixed
sibling_count5 0.004279 0.02475 0.1729 -0.04423 0.05279 fixed
sibling_count5+ 0.009765 0.02168 0.4505 -0.03272 0.05225 fixed
sd_(Intercept).mother_pidlink 0.1641 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3415 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.2047 0.248 0.8255 -0.2814 0.6908 fixed
birth_order 0.003732 0.00347 1.076 -0.003069 0.01053 fixed
poly(age, 3, raw = TRUE)1 -0.006952 0.02663 -0.2611 -0.05914 0.04524 fixed
poly(age, 3, raw = TRUE)2 0.0002152 0.0009121 0.236 -0.001572 0.002003 fixed
poly(age, 3, raw = TRUE)3 -0.00000108 0.000009998 -0.108 -0.00002068 0.00001852 fixed
male 0.0186 0.0125 1.488 -0.005904 0.0431 fixed
sibling_count3 -0.0247 0.02323 -1.063 -0.07023 0.02084 fixed
sibling_count4 -0.008205 0.02369 -0.3464 -0.05464 0.03823 fixed
sibling_count5 -0.002088 0.02545 -0.08207 -0.05196 0.04778 fixed
sibling_count5+ -0.003568 0.02497 -0.1429 -0.05251 0.04537 fixed
sd_(Intercept).mother_pidlink 0.1639 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3416 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.196 0.2482 0.7895 -0.2905 0.6824 fixed
poly(age, 3, raw = TRUE)1 -0.006367 0.02663 -0.2391 -0.05856 0.04583 fixed
poly(age, 3, raw = TRUE)2 0.0001927 0.0009122 0.2113 -0.001595 0.001981 fixed
poly(age, 3, raw = TRUE)3 -0.0000008051 0.00001 -0.0805 -0.00002041 0.0000188 fixed
male 0.01888 0.0125 1.51 -0.005619 0.04339 fixed
sibling_count3 -0.03091 0.02358 -1.311 -0.07713 0.0153 fixed
sibling_count4 -0.01824 0.02432 -0.7502 -0.0659 0.02942 fixed
sibling_count5 -0.008351 0.02645 -0.3157 -0.0602 0.04349 fixed
sibling_count5+ -0.00879 0.02563 -0.343 -0.05902 0.04144 fixed
birth_order_nonlinear2 0.02643 0.01663 1.589 -0.006167 0.05903 fixed
birth_order_nonlinear3 0.03556 0.01942 1.831 -0.002503 0.07363 fixed
birth_order_nonlinear4 0.04224 0.02305 1.833 -0.002935 0.08741 fixed
birth_order_nonlinear5 0.00225 0.02807 0.08016 -0.05276 0.05726 fixed
birth_order_nonlinear5+ 0.03512 0.02623 1.339 -0.01628 0.08652 fixed
sd_(Intercept).mother_pidlink 0.1636 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3417 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.1632 0.2485 0.6567 -0.3239 0.6503 fixed
poly(age, 3, raw = TRUE)1 -0.002079 0.02667 -0.07794 -0.05436 0.0502 fixed
poly(age, 3, raw = TRUE)2 0.00001812 0.000914 0.01983 -0.001773 0.00181 fixed
poly(age, 3, raw = TRUE)3 0.000001353 0.00001002 0.135 -0.00001829 0.000021 fixed
male 0.01964 0.01251 1.57 -0.004879 0.04415 fixed
count_birth_order2/2 0.02775 0.03526 0.7871 -0.04135 0.09686 fixed
count_birth_order1/3 -0.01368 0.03019 -0.453 -0.07286 0.0455 fixed
count_birth_order2/3 -0.02855 0.03284 -0.8694 -0.09292 0.03582 fixed
count_birth_order3/3 0.007524 0.03635 0.207 -0.06372 0.07877 fixed
count_birth_order1/4 -0.02249 0.03274 -0.6868 -0.08666 0.04169 fixed
count_birth_order2/4 -0.01115 0.03446 -0.3235 -0.07868 0.05639 fixed
count_birth_order3/4 0.08672 0.03724 2.328 0.01372 0.1597 fixed
count_birth_order4/4 -0.01743 0.04042 -0.4312 -0.09665 0.06179 fixed
count_birth_order1/5 0.01313 0.03952 0.3322 -0.06433 0.09059 fixed
count_birth_order2/5 0.02348 0.04048 0.58 -0.05586 0.1028 fixed
count_birth_order3/5 0.01464 0.0416 0.3519 -0.06689 0.09617 fixed
count_birth_order4/5 0.04861 0.04232 1.149 -0.03434 0.1316 fixed
count_birth_order5/5 -0.036 0.04295 -0.8381 -0.1202 0.04819 fixed
count_birth_order1/5+ -0.04299 0.03567 -1.205 -0.1129 0.02692 fixed
count_birth_order2/5+ 0.07643 0.03759 2.034 0.002765 0.1501 fixed
count_birth_order3/5+ -0.02567 0.03621 -0.7089 -0.09664 0.0453 fixed
count_birth_order4/5+ 0.05458 0.03517 1.552 -0.01435 0.1235 fixed
count_birth_order5/5+ 0.0148 0.03721 0.3979 -0.05812 0.08773 fixed
count_birth_order5+/5+ 0.0284 0.02839 1 -0.02725 0.08405 fixed
sd_(Intercept).mother_pidlink 0.1614 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.342 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 3280 3348 -1629 3258 NA NA NA
12 3280 3355 -1628 3256 1.161 1 0.2813
16 3283 3382 -1625 3251 5.449 4 0.2443
26 3282 3444 -1615 3230 20.81 10 0.02245

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.1848 0.2525 0.732 -0.3101 0.6797 fixed
poly(age, 3, raw = TRUE)1 -0.004672 0.02712 -0.1723 -0.05782 0.04848 fixed
poly(age, 3, raw = TRUE)2 0.0001606 0.0009297 0.1727 -0.001662 0.001983 fixed
poly(age, 3, raw = TRUE)3 -0.0000009046 0.0000102 -0.08866 -0.0000209 0.00001909 fixed
male 0.01747 0.01271 1.375 -0.007427 0.04238 fixed
sibling_count3 -0.006028 0.0208 -0.2897 -0.0468 0.03475 fixed
sibling_count4 -0.008332 0.02186 -0.3811 -0.05118 0.03451 fixed
sibling_count5 0.00707 0.02522 0.2804 -0.04235 0.05649 fixed
sibling_count5+ 0.01936 0.02141 0.9043 -0.02261 0.06133 fixed
sd_(Intercept).mother_pidlink 0.1663 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3423 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.1905 0.2526 0.7539 -0.3047 0.6856 fixed
birth_order 0.003283 0.004052 0.81 -0.00466 0.01123 fixed
poly(age, 3, raw = TRUE)1 -0.005701 0.02715 -0.21 -0.05891 0.04751 fixed
poly(age, 3, raw = TRUE)2 0.0001888 0.0009304 0.2029 -0.001635 0.002012 fixed
poly(age, 3, raw = TRUE)3 -0.000001065 0.00001021 -0.1043 -0.00002107 0.00001894 fixed
male 0.01743 0.01271 1.372 -0.007468 0.04234 fixed
sibling_count3 -0.007723 0.02091 -0.3694 -0.04871 0.03326 fixed
sibling_count4 -0.01195 0.02231 -0.5356 -0.05569 0.03178 fixed
sibling_count5 0.00124 0.02622 0.04729 -0.05016 0.05264 fixed
sibling_count5+ 0.007432 0.02599 0.2859 -0.04351 0.05837 fixed
sd_(Intercept).mother_pidlink 0.1663 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3424 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.1896 0.2527 0.7506 -0.3056 0.6848 fixed
poly(age, 3, raw = TRUE)1 -0.006112 0.02713 -0.2253 -0.05929 0.04707 fixed
poly(age, 3, raw = TRUE)2 0.0001968 0.0009299 0.2116 -0.001626 0.002019 fixed
poly(age, 3, raw = TRUE)3 -0.000001088 0.0000102 -0.1067 -0.00002109 0.00001891 fixed
male 0.01806 0.0127 1.422 -0.006835 0.04296 fixed
sibling_count3 -0.01698 0.02128 -0.7977 -0.05869 0.02473 fixed
sibling_count4 -0.0222 0.02304 -0.9634 -0.06736 0.02296 fixed
sibling_count5 -0.001415 0.0272 -0.052 -0.05473 0.0519 fixed
sibling_count5+ 0.001559 0.02661 0.0586 -0.0506 0.05372 fixed
birth_order_nonlinear2 0.03129 0.0164 1.908 -0.000851 0.06344 fixed
birth_order_nonlinear3 0.04782 0.01938 2.467 0.00983 0.0858 fixed
birth_order_nonlinear4 0.02218 0.02437 0.9103 -0.02557 0.06994 fixed
birth_order_nonlinear5 -0.006486 0.03001 -0.2161 -0.06531 0.05234 fixed
birth_order_nonlinear5+ 0.04426 0.02993 1.479 -0.0144 0.1029 fixed
sd_(Intercept).mother_pidlink 0.1653 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3425 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.1681 0.2535 0.663 -0.3288 0.665 fixed
poly(age, 3, raw = TRUE)1 -0.003336 0.02724 -0.1225 -0.05672 0.05005 fixed
poly(age, 3, raw = TRUE)2 0.0000858 0.0009339 0.09188 -0.001745 0.001916 fixed
poly(age, 3, raw = TRUE)3 0.0000002996 0.00001025 0.02923 -0.00001979 0.00002039 fixed
male 0.01896 0.01272 1.49 -0.005974 0.04389 fixed
count_birth_order2/2 0.02905 0.03134 0.9271 -0.03237 0.09047 fixed
count_birth_order1/3 -0.002032 0.02696 -0.07537 -0.05487 0.05081 fixed
count_birth_order2/3 -0.007826 0.02992 -0.2615 -0.06648 0.05082 fixed
count_birth_order3/3 0.02743 0.03219 0.8522 -0.03565 0.09051 fixed
count_birth_order1/4 -0.02844 0.03159 -0.9001 -0.09036 0.03349 fixed
count_birth_order2/4 0.01055 0.03314 0.3183 -0.0544 0.07549 fixed
count_birth_order3/4 0.05262 0.03435 1.532 -0.01471 0.12 fixed
count_birth_order4/4 -0.02922 0.0373 -0.7834 -0.1023 0.04389 fixed
count_birth_order1/5 -0.005244 0.04125 -0.1271 -0.0861 0.07561 fixed
count_birth_order2/5 0.01438 0.04592 0.3131 -0.07563 0.1044 fixed
count_birth_order3/5 0.03522 0.04477 0.7867 -0.05253 0.123 fixed
count_birth_order4/5 0.06448 0.04311 1.496 -0.02001 0.149 fixed
count_birth_order5/5 -0.02928 0.04577 -0.6397 -0.119 0.06042 fixed
count_birth_order1/5+ -0.03966 0.04016 -0.9875 -0.1184 0.03906 fixed
count_birth_order2/5+ 0.09011 0.04053 2.224 0.01068 0.1695 fixed
count_birth_order3/5+ 0.02288 0.0387 0.5913 -0.05297 0.09873 fixed
count_birth_order4/5+ 0.02081 0.03882 0.5362 -0.05527 0.09689 fixed
count_birth_order5/5+ 0.005412 0.03647 0.1484 -0.06607 0.07689 fixed
count_birth_order5+/5+ 0.04579 0.02876 1.592 -0.01057 0.1022 fixed
sd_(Intercept).mother_pidlink 0.1631 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3434 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 3231 3300 -1605 3209 NA NA NA
12 3233 3307 -1604 3209 0.6582 1 0.4172
16 3232 3331 -1600 3200 8.947 4 0.06245
26 3241 3402 -1595 3189 10.51 10 0.3971

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Sector_Construction

birthorder <- birthorder %>% mutate(outcome = `Sector_Construction`)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.00014 0.007992 -0.01751 -0.0158 0.01552 fixed
poly(age, 3, raw = TRUE)1 0.0001646 0.0007143 0.2304 -0.001235 0.001565 fixed
poly(age, 3, raw = TRUE)2 -0.000005987 0.00001998 -0.2997 -0.00004514 0.00003317 fixed
poly(age, 3, raw = TRUE)3 0.00000005958 0.0000001767 0.3371 -0.0000002868 0.000000406 fixed
male 0.001437 0.0007496 1.917 -0.00003186 0.002907 fixed
sibling_count3 -0.001779 0.001507 -1.18 -0.004733 0.001175 fixed
sibling_count4 0.0006745 0.001514 0.4455 -0.002293 0.003642 fixed
sibling_count5 -0.0008865 0.001566 -0.5663 -0.003955 0.002182 fixed
sibling_count5+ -0.0004652 0.001227 -0.3793 -0.002869 0.001939 fixed
sd_(Intercept).mother_pidlink 0.000000004242 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.03649 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.0001227 0.007992 -0.01535 -0.01579 0.01554 fixed
birth_order 0.00008633 0.0001468 0.5883 -0.0002013 0.000374 fixed
poly(age, 3, raw = TRUE)1 0.0001414 0.0007155 0.1976 -0.001261 0.001544 fixed
poly(age, 3, raw = TRUE)2 -0.00000519 0.00002002 -0.2592 -0.00004443 0.00003405 fixed
poly(age, 3, raw = TRUE)3 0.00000005258 0.0000001771 0.2968 -0.0000002946 0.0000003998 fixed
male 0.001435 0.0007497 1.914 -0.00003438 0.002904 fixed
sibling_count3 -0.001804 0.001508 -1.197 -0.00476 0.001151 fixed
sibling_count4 0.000616 0.001517 0.4059 -0.002358 0.00359 fixed
sibling_count5 -0.0009842 0.001574 -0.6252 -0.00407 0.002101 fixed
sibling_count5+ -0.0007791 0.001338 -0.5825 -0.003401 0.001843 fixed
sd_(Intercept).mother_pidlink 0.000000005156 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.03649 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.0006085 0.008007 -0.07599 -0.0163 0.01508 fixed
poly(age, 3, raw = TRUE)1 0.0002089 0.0007162 0.2916 -0.001195 0.001613 fixed
poly(age, 3, raw = TRUE)2 -0.000007051 0.00002004 -0.3519 -0.00004633 0.00003222 fixed
poly(age, 3, raw = TRUE)3 0.00000006636 0.0000001772 0.3744 -0.000000281 0.0000004138 fixed
male 0.00143 0.0007498 1.907 -0.00003985 0.002899 fixed
sibling_count3 -0.001577 0.001534 -1.028 -0.004583 0.001429 fixed
sibling_count4 0.001223 0.001566 0.7812 -0.001846 0.004293 fixed
sibling_count5 -0.0005539 0.001638 -0.3382 -0.003764 0.002656 fixed
sibling_count5+ -0.00006166 0.00141 -0.04372 -0.002826 0.002702 fixed
birth_order_nonlinear2 -0.00004771 0.001109 -0.04303 -0.002221 0.002126 fixed
birth_order_nonlinear3 -0.0008901 0.001289 -0.6904 -0.003417 0.001637 fixed
birth_order_nonlinear4 -0.002 0.001448 -1.382 -0.004838 0.0008371 fixed
birth_order_nonlinear5 0.000813 0.001625 0.5004 -0.002372 0.003998 fixed
birth_order_nonlinear5+ -0.0004879 0.001322 -0.369 -0.003079 0.002104 fixed
sd_(Intercept).mother_pidlink 0.00000001036 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.03649 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.001129 0.00804 -0.1404 -0.01689 0.01463 fixed
poly(age, 3, raw = TRUE)1 0.0002277 0.0007171 0.3175 -0.001178 0.001633 fixed
poly(age, 3, raw = TRUE)2 -0.000007642 0.00002005 -0.3811 -0.00004695 0.00003166 fixed
poly(age, 3, raw = TRUE)3 0.00000007203 0.0000001773 0.4062 -0.0000002756 0.0000004196 fixed
male 0.001399 0.0007502 1.865 -0.00007127 0.00287 fixed
count_birth_order2/2 0.0008908 0.002215 0.4021 -0.003451 0.005233 fixed
count_birth_order1/3 -0.001422 0.002092 -0.6798 -0.005522 0.002678 fixed
count_birth_order2/3 -0.001456 0.002328 -0.6254 -0.006017 0.003106 fixed
count_birth_order3/3 -0.001499 0.002556 -0.5867 -0.006508 0.003509 fixed
count_birth_order1/4 0.003658 0.002303 1.588 -0.0008556 0.008173 fixed
count_birth_order2/4 0.001647 0.00247 0.6669 -0.003194 0.006489 fixed
count_birth_order3/4 -0.001469 0.002605 -0.5637 -0.006575 0.003638 fixed
count_birth_order4/4 -0.001533 0.002805 -0.5465 -0.00703 0.003965 fixed
count_birth_order1/5 -0.001484 0.002599 -0.5709 -0.006577 0.00361 fixed
count_birth_order2/5 0.00277 0.002759 1.004 -0.002637 0.008178 fixed
count_birth_order3/5 -0.001468 0.002903 -0.5055 -0.007158 0.004222 fixed
count_birth_order4/5 -0.001597 0.003066 -0.521 -0.007606 0.004411 fixed
count_birth_order5/5 -0.001528 0.00307 -0.4976 -0.007545 0.00449 fixed
count_birth_order1/5+ 0.0001466 0.00201 0.07294 -0.003793 0.004087 fixed
count_birth_order2/5+ -0.001418 0.00209 -0.6783 -0.005515 0.002679 fixed
count_birth_order3/5+ 0.0002647 0.002069 0.1279 -0.003791 0.00432 fixed
count_birth_order4/5+ -0.001477 0.002027 -0.729 -0.00545 0.002495 fixed
count_birth_order5/5+ 0.001744 0.002049 0.8512 -0.002272 0.00576 fixed
count_birth_order5+/5+ -0.0001939 0.001632 -0.1188 -0.003393 0.003005 fixed
sd_(Intercept).mother_pidlink 0.00000001384 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.03649 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -36894 -36815 18458 -36916 NA NA NA
12 -36892 -36806 18458 -36916 0.3464 1 0.5562
16 -36887 -36772 18460 -36919 3.056 4 0.5484
26 -36874 -36687 18463 -36926 7.077 10 0.7181

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.01904 0.02438 -0.7808 -0.06683 0.02875 fixed
poly(age, 3, raw = TRUE)1 0.00222 0.002615 0.8488 -0.002906 0.007345 fixed
poly(age, 3, raw = TRUE)2 -0.00006986 0.00008957 -0.7799 -0.0002454 0.0001057 fixed
poly(age, 3, raw = TRUE)3 0.0000006769 0.0000009816 0.6896 -0.000001247 0.000002601 fixed
male 0.001275 0.001234 1.033 -0.001144 0.003695 fixed
sibling_count3 -0.002135 0.001996 -1.07 -0.006047 0.001777 fixed
sibling_count4 -0.003247 0.002052 -1.582 -0.007268 0.0007746 fixed
sibling_count5 -0.001205 0.002273 -0.5304 -0.00566 0.003249 fixed
sibling_count5+ -0.002255 0.00196 -1.15 -0.006098 0.001587 fixed
sd_(Intercept).mother_pidlink 0 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.03688 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.01893 0.02439 -0.7762 -0.06673 0.02887 fixed
birth_order 0.0001139 0.0003773 0.302 -0.0006255 0.0008534 fixed
poly(age, 3, raw = TRUE)1 0.002195 0.002617 0.839 -0.002933 0.007324 fixed
poly(age, 3, raw = TRUE)2 -0.00006929 0.0000896 -0.7733 -0.0002449 0.0001063 fixed
poly(age, 3, raw = TRUE)3 0.0000006754 0.0000009817 0.688 -0.000001249 0.000002599 fixed
male 0.001271 0.001235 1.029 -0.001149 0.00369 fixed
sibling_count3 -0.002192 0.002005 -1.093 -0.006121 0.001738 fixed
sibling_count4 -0.00337 0.002092 -1.611 -0.007472 0.0007308 fixed
sibling_count5 -0.001409 0.002371 -0.5944 -0.006056 0.003238 fixed
sibling_count5+ -0.00266 0.002374 -1.12 -0.007313 0.001994 fixed
sd_(Intercept).mother_pidlink 0 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.03689 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.01804 0.0244 -0.7394 -0.06587 0.02979 fixed
poly(age, 3, raw = TRUE)1 0.002198 0.002617 0.8399 -0.002931 0.007327 fixed
poly(age, 3, raw = TRUE)2 -0.00006814 0.00008961 -0.7603 -0.0002438 0.0001075 fixed
poly(age, 3, raw = TRUE)3 0.0000006483 0.0000009821 0.6601 -0.000001276 0.000002573 fixed
male 0.001222 0.001235 0.9901 -0.001198 0.003642 fixed
sibling_count3 -0.001531 0.002047 -0.7481 -0.005542 0.00248 fixed
sibling_count4 -0.002233 0.002175 -1.027 -0.006497 0.00203 fixed
sibling_count5 0.0001622 0.002493 0.06504 -0.004725 0.005049 fixed
sibling_count5+ -0.001611 0.002443 -0.6594 -0.0064 0.003178 fixed
birth_order_nonlinear2 -0.003122 0.001656 -1.886 -0.006368 0.0001233 fixed
birth_order_nonlinear3 -0.002794 0.00194 -1.44 -0.006597 0.001009 fixed
birth_order_nonlinear4 -0.003035 0.002365 -1.284 -0.00767 0.0016 fixed
birth_order_nonlinear5 -0.003377 0.002899 -1.165 -0.009059 0.002304 fixed
birth_order_nonlinear5+ -0.0000936 0.002817 -0.03323 -0.005614 0.005427 fixed
sd_(Intercept).mother_pidlink 0.00000001012 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.03688 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.01888 0.02451 -0.7703 -0.06692 0.02916 fixed
poly(age, 3, raw = TRUE)1 0.002348 0.002629 0.8932 -0.002804 0.0075 fixed
poly(age, 3, raw = TRUE)2 -0.00007276 0.00009004 -0.8081 -0.0002492 0.0001037 fixed
poly(age, 3, raw = TRUE)3 0.0000006932 0.0000009869 0.7024 -0.000001241 0.000002627 fixed
male 0.001217 0.001237 0.9839 -0.001208 0.003642 fixed
count_birth_order2/2 -0.005185 0.003234 -1.603 -0.01152 0.001153 fixed
count_birth_order1/3 -0.002396 0.002674 -0.8963 -0.007636 0.002844 fixed
count_birth_order2/3 -0.005014 0.002973 -1.687 -0.01084 0.0008128 fixed
count_birth_order3/3 -0.005052 0.00324 -1.559 -0.0114 0.001299 fixed
count_birth_order1/4 -0.004789 0.003067 -1.561 -0.0108 0.001222 fixed
count_birth_order2/4 -0.004976 0.003253 -1.529 -0.01135 0.0014 fixed
count_birth_order3/4 -0.004932 0.003375 -1.461 -0.01155 0.001683 fixed
count_birth_order4/4 -0.005233 0.003613 -1.448 -0.01231 0.001848 fixed
count_birth_order1/5 0.004607 0.004041 1.14 -0.003313 0.01253 fixed
count_birth_order2/5 -0.004763 0.004362 -1.092 -0.01331 0.003786 fixed
count_birth_order3/5 -0.004893 0.004147 -1.18 -0.01302 0.003235 fixed
count_birth_order4/5 -0.005202 0.004046 -1.286 -0.01313 0.002729 fixed
count_birth_order5/5 -0.005179 0.004259 -1.216 -0.01353 0.003169 fixed
count_birth_order1/5+ -0.004681 0.003823 -1.224 -0.01217 0.002812 fixed
count_birth_order2/5+ -0.00473 0.003857 -1.226 -0.01229 0.002829 fixed
count_birth_order3/5+ -0.004821 0.003765 -1.28 -0.0122 0.00256 fixed
count_birth_order4/5+ -0.004809 0.003688 -1.304 -0.01204 0.002419 fixed
count_birth_order5/5+ -0.004901 0.003524 -1.391 -0.01181 0.002007 fixed
count_birth_order5+/5+ -0.002395 0.002722 -0.8798 -0.007729 0.00294 fixed
sd_(Intercept).mother_pidlink 0.0000000003251 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.03691 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -13809 -13741 6916 -13831 NA NA NA
12 -13807 -13733 6916 -13831 0.09146 1 0.7623
16 -13805 -13705 6918 -13837 5.574 4 0.2333
26 -13790 -13628 6921 -13842 4.71 10 0.9097

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.01904 0.02419 -0.7868 -0.06646 0.02838 fixed
poly(age, 3, raw = TRUE)1 0.002273 0.002597 0.875 -0.002818 0.007363 fixed
poly(age, 3, raw = TRUE)2 -0.00007096 0.00008898 -0.7975 -0.0002454 0.0001034 fixed
poly(age, 3, raw = TRUE)3 0.0000006852 0.0000009753 0.7025 -0.000001226 0.000002597 fixed
male 0.001309 0.001226 1.067 -0.001094 0.003712 fixed
sibling_count3 -0.002695 0.002183 -1.235 -0.006973 0.001583 fixed
sibling_count4 -0.004092 0.002196 -1.864 -0.008396 0.0002107 fixed
sibling_count5 -0.002261 0.002303 -0.9815 -0.006775 0.002254 fixed
sibling_count5+ -0.00326 0.00202 -1.614 -0.00722 0.0006995 fixed
sd_(Intercept).mother_pidlink 0.00000001331 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.03676 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.01896 0.0242 -0.7833 -0.06639 0.02847 fixed
birth_order 0.0001029 0.0003309 0.3109 -0.0005457 0.0007515 fixed
poly(age, 3, raw = TRUE)1 0.002253 0.002598 0.8669 -0.00284 0.007346 fixed
poly(age, 3, raw = TRUE)2 -0.00007051 0.000089 -0.7923 -0.000245 0.0001039 fixed
poly(age, 3, raw = TRUE)3 0.0000006846 0.0000009754 0.7018 -0.000001227 0.000002596 fixed
male 0.001304 0.001226 1.064 -0.001099 0.003708 fixed
sibling_count3 -0.002747 0.002189 -1.255 -0.007037 0.001544 fixed
sibling_count4 -0.004195 0.00222 -1.889 -0.008547 0.0001569 fixed
sibling_count5 -0.002432 0.002369 -1.027 -0.007075 0.00221 fixed
sibling_count5+ -0.003621 0.00233 -1.554 -0.008187 0.0009452 fixed
sd_(Intercept).mother_pidlink 0.00000003172 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.03676 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.01787 0.02422 -0.7379 -0.06533 0.02959 fixed
poly(age, 3, raw = TRUE)1 0.002244 0.002599 0.8634 -0.00285 0.007338 fixed
poly(age, 3, raw = TRUE)2 -0.00006906 0.00008903 -0.7757 -0.0002435 0.0001054 fixed
poly(age, 3, raw = TRUE)3 0.0000006551 0.0000009758 0.6713 -0.000001258 0.000002568 fixed
male 0.001255 0.001226 1.023 -0.001149 0.003658 fixed
sibling_count3 -0.002103 0.002228 -0.944 -0.006471 0.002264 fixed
sibling_count4 -0.003207 0.00229 -1.4 -0.007695 0.001281 fixed
sibling_count5 -0.001055 0.002481 -0.4253 -0.005918 0.003807 fixed
sibling_count5+ -0.002713 0.002404 -1.129 -0.007424 0.001998 fixed
birth_order_nonlinear2 -0.003291 0.001669 -1.972 -0.006562 -0.00001936 fixed
birth_order_nonlinear3 -0.002759 0.001942 -1.42 -0.006566 0.001048 fixed
birth_order_nonlinear4 -0.002861 0.002297 -1.245 -0.007363 0.001641 fixed
birth_order_nonlinear5 -0.003159 0.002798 -1.129 -0.008642 0.002324 fixed
birth_order_nonlinear5+ -0.00035 0.002537 -0.1379 -0.005323 0.004623 fixed
sd_(Intercept).mother_pidlink 0.000000005965 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.03675 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.01698 0.02429 -0.6991 -0.0646 0.03063 fixed
poly(age, 3, raw = TRUE)1 0.00227 0.002608 0.8703 -0.002841 0.007381 fixed
poly(age, 3, raw = TRUE)2 -0.0000702 0.00008935 -0.7857 -0.0002453 0.0001049 fixed
poly(age, 3, raw = TRUE)3 0.0000006703 0.0000009796 0.6843 -0.00000125 0.00000259 fixed
male 0.001227 0.001229 0.9981 -0.001183 0.003637 fixed
count_birth_order2/2 -0.006385 0.00353 -1.809 -0.0133 0.0005343 fixed
count_birth_order1/3 -0.003068 0.002946 -1.041 -0.008843 0.002707 fixed
count_birth_order2/3 -0.006267 0.00321 -1.952 -0.01256 0.0000252 fixed
count_birth_order3/3 -0.006277 0.003561 -1.763 -0.01326 0.0007021 fixed
count_birth_order1/4 -0.006116 0.003198 -1.913 -0.01238 0.0001512 fixed
count_birth_order2/4 -0.006265 0.003371 -1.859 -0.01287 0.0003419 fixed
count_birth_order3/4 -0.006214 0.003651 -1.702 -0.01337 0.0009411 fixed
count_birth_order4/4 -0.006527 0.003966 -1.646 -0.0143 0.001246 fixed
count_birth_order1/5 0.001582 0.003866 0.4093 -0.005996 0.00916 fixed
count_birth_order2/5 -0.006089 0.003966 -1.535 -0.01386 0.001684 fixed
count_birth_order3/5 -0.006068 0.004082 -1.486 -0.01407 0.001933 fixed
count_birth_order4/5 -0.006315 0.004157 -1.519 -0.01446 0.001832 fixed
count_birth_order5/5 -0.006395 0.004222 -1.515 -0.01467 0.001879 fixed
count_birth_order1/5+ -0.00598 0.003485 -1.716 -0.01281 0.0008499 fixed
count_birth_order2/5+ -0.006117 0.003686 -1.66 -0.01334 0.001106 fixed
count_birth_order3/5+ -0.006069 0.003552 -1.709 -0.01303 0.0008932 fixed
count_birth_order4/5+ -0.006211 0.003452 -1.799 -0.01298 0.0005554 fixed
count_birth_order5/5+ -0.006157 0.003658 -1.683 -0.01333 0.001011 fixed
count_birth_order5+/5+ -0.004098 0.002734 -1.499 -0.009457 0.00126 fixed
sd_(Intercept).mother_pidlink 0.00000001386 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.03678 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -13925 -13856 6973 -13947 NA NA NA
12 -13923 -13848 6973 -13947 0.09692 1 0.7556
16 -13920 -13821 6976 -13952 5.536 4 0.2366
26 -13904 -13743 6978 -13956 4.304 10 0.9326

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.02014 0.02484 -0.8108 -0.06881 0.02854 fixed
poly(age, 3, raw = TRUE)1 0.002324 0.002668 0.871 -0.002905 0.007553 fixed
poly(age, 3, raw = TRUE)2 -0.00007334 0.00009147 -0.8018 -0.0002526 0.0001059 fixed
poly(age, 3, raw = TRUE)3 0.0000007122 0.000001003 0.7098 -0.000001255 0.000002679 fixed
male 0.001306 0.001257 1.039 -0.001157 0.003769 fixed
sibling_count3 -0.002062 0.001971 -1.046 -0.005924 0.0018 fixed
sibling_count4 -0.003122 0.002055 -1.52 -0.00715 0.000905 fixed
sibling_count5 -0.0008557 0.002335 -0.3665 -0.005432 0.003721 fixed
sibling_count5+ -0.002037 0.001982 -1.028 -0.005921 0.001848 fixed
sd_(Intercept).mother_pidlink 0 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.03721 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.01997 0.02484 -0.804 -0.06867 0.02872 fixed
birth_order 0.0001248 0.0003922 0.3183 -0.0006439 0.0008935 fixed
poly(age, 3, raw = TRUE)1 0.002292 0.00267 0.8585 -0.002941 0.007526 fixed
poly(age, 3, raw = TRUE)2 -0.00007255 0.00009151 -0.7927 -0.0002519 0.0001068 fixed
poly(age, 3, raw = TRUE)3 0.0000007088 0.000001004 0.7062 -0.000001258 0.000002676 fixed
male 0.001304 0.001257 1.038 -0.001159 0.003767 fixed
sibling_count3 -0.002125 0.001981 -1.073 -0.006007 0.001757 fixed
sibling_count4 -0.003256 0.002098 -1.552 -0.007368 0.0008552 fixed
sibling_count5 -0.001071 0.002431 -0.4406 -0.005836 0.003694 fixed
sibling_count5+ -0.00248 0.002422 -1.024 -0.007228 0.002268 fixed
sd_(Intercept).mother_pidlink 0 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.03721 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.01926 0.02486 -0.7748 -0.06798 0.02946 fixed
poly(age, 3, raw = TRUE)1 0.002317 0.00267 0.8677 -0.002916 0.007551 fixed
poly(age, 3, raw = TRUE)2 -0.00007222 0.00009152 -0.7891 -0.0002516 0.0001072 fixed
poly(age, 3, raw = TRUE)3 0.000000692 0.000001004 0.6892 -0.000001276 0.00000266 fixed
male 0.001266 0.001257 1.008 -0.001197 0.00373 fixed
sibling_count3 -0.00148 0.002025 -0.7309 -0.005448 0.002488 fixed
sibling_count4 -0.002126 0.002182 -0.9744 -0.006403 0.002151 fixed
sibling_count5 0.0004233 0.002547 0.1662 -0.004569 0.005415 fixed
sibling_count5+ -0.001472 0.002497 -0.5893 -0.006366 0.003423 fixed
birth_order_nonlinear2 -0.003103 0.001663 -1.866 -0.006363 0.0001568 fixed
birth_order_nonlinear3 -0.002762 0.00196 -1.409 -0.006603 0.00108 fixed
birth_order_nonlinear4 -0.003046 0.002457 -1.239 -0.007861 0.00177 fixed
birth_order_nonlinear5 -0.003403 0.003029 -1.123 -0.009339 0.002533 fixed
birth_order_nonlinear5+ 0.00008479 0.002932 0.02891 -0.005663 0.005832 fixed
sd_(Intercept).mother_pidlink 0 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.03721 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.01999 0.02495 -0.8009 -0.0689 0.02892 fixed
poly(age, 3, raw = TRUE)1 0.002455 0.002681 0.9155 -0.0028 0.00771 fixed
poly(age, 3, raw = TRUE)2 -0.00007656 0.00009193 -0.8328 -0.0002567 0.0001036 fixed
poly(age, 3, raw = TRUE)3 0.0000007349 0.000001009 0.7286 -0.000001242 0.000002712 fixed
male 0.00125 0.001259 0.9929 -0.001218 0.003719 fixed
count_birth_order2/2 -0.004988 0.003173 -1.572 -0.01121 0.001231 fixed
count_birth_order1/3 -0.002318 0.002649 -0.8749 -0.00751 0.002875 fixed
count_birth_order2/3 -0.00488 0.002948 -1.655 -0.01066 0.0008983 fixed
count_birth_order3/3 -0.004906 0.003177 -1.544 -0.01113 0.001321 fixed
count_birth_order1/4 -0.004633 0.00311 -1.49 -0.01073 0.001461 fixed
count_birth_order2/4 -0.004836 0.003269 -1.479 -0.01124 0.001571 fixed
count_birth_order3/4 -0.004774 0.003394 -1.407 -0.01142 0.001878 fixed
count_birth_order4/4 -0.005074 0.003687 -1.376 -0.0123 0.002153 fixed
count_birth_order1/5 0.00475 0.004066 1.168 -0.003218 0.01272 fixed
count_birth_order2/5 -0.004579 0.004544 -1.008 -0.01349 0.004327 fixed
count_birth_order3/5 -0.004708 0.004432 -1.062 -0.0134 0.003979 fixed
count_birth_order4/5 -0.005011 0.004267 -1.174 -0.01337 0.003352 fixed
count_birth_order5/5 -0.005039 0.004533 -1.111 -0.01392 0.003847 fixed
count_birth_order1/5+ -0.004509 0.003959 -1.139 -0.01227 0.003251 fixed
count_birth_order2/5+ -0.004502 0.004007 -1.124 -0.01236 0.003352 fixed
count_birth_order3/5+ -0.00464 0.003827 -1.212 -0.01214 0.002862 fixed
count_birth_order4/5+ -0.00463 0.003846 -1.204 -0.01217 0.002908 fixed
count_birth_order5/5+ -0.00472 0.003611 -1.307 -0.0118 0.002357 fixed
count_birth_order5+/5+ -0.002027 0.002777 -0.7299 -0.007471 0.003416 fixed
sd_(Intercept).mother_pidlink 0 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.03723 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -13505 -13437 6763 -13527 NA NA NA
12 -13503 -13429 6763 -13527 0.1016 1 0.75
16 -13500 -13401 6766 -13532 5.511 4 0.2388
26 -13485 -13324 6768 -13537 4.409 10 0.927

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Sector_Electricity, gas, water

birthorder <- birthorder %>% mutate(outcome = `Sector_Electricity, gas, water`)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.01323 0.0289 0.4577 -0.04342 0.06987 fixed
poly(age, 3, raw = TRUE)1 0.0002009 0.002584 0.07776 -0.004864 0.005266 fixed
poly(age, 3, raw = TRUE)2 0.00001341 0.0000723 0.1854 -0.0001283 0.0001551 fixed
poly(age, 3, raw = TRUE)3 -0.0000002622 0.0000006399 -0.4098 -0.000001516 0.0000009919 fixed
male 0.00005814 0.002704 0.02151 -0.005241 0.005357 fixed
sibling_count3 -0.006514 0.005475 -1.19 -0.01724 0.004217 fixed
sibling_count4 0.000643 0.005507 0.1168 -0.01015 0.01144 fixed
sibling_count5 -0.01138 0.005702 -1.996 -0.02256 -0.000205 fixed
sibling_count5+ -0.007491 0.004458 -1.68 -0.01623 0.001247 fixed
sd_(Intercept).mother_pidlink 0.02111 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1299 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.01323 0.0289 0.4578 -0.04341 0.06988 fixed
birth_order 0.00004662 0.0005342 0.08727 -0.001 0.001094 fixed
poly(age, 3, raw = TRUE)1 0.0001884 0.002588 0.07277 -0.004885 0.005262 fixed
poly(age, 3, raw = TRUE)2 0.00001385 0.00007248 0.191 -0.0001282 0.0001559 fixed
poly(age, 3, raw = TRUE)3 -0.0000002661 0.0000006415 -0.4149 -0.000001523 0.0000009911 fixed
male 0.00005683 0.002704 0.02102 -0.005243 0.005356 fixed
sibling_count3 -0.006527 0.005477 -1.192 -0.01726 0.004208 fixed
sibling_count4 0.000612 0.005519 0.1109 -0.01021 0.01143 fixed
sibling_count5 -0.01143 0.005733 -1.994 -0.02267 -0.0001956 fixed
sibling_count5+ -0.007659 0.004856 -1.577 -0.01718 0.001858 fixed
sd_(Intercept).mother_pidlink 0.02113 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1299 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.01097 0.02895 0.3791 -0.04577 0.06772 fixed
poly(age, 3, raw = TRUE)1 0.0002737 0.002591 0.1057 -0.004805 0.005352 fixed
poly(age, 3, raw = TRUE)2 0.000009057 0.00007253 0.1249 -0.0001331 0.0001512 fixed
poly(age, 3, raw = TRUE)3 -0.0000002072 0.0000006418 -0.3228 -0.000001465 0.000001051 fixed
male 0.00005964 0.002704 0.02205 -0.00524 0.00536 fixed
sibling_count3 -0.006954 0.005568 -1.249 -0.01787 0.003959 fixed
sibling_count4 0.0003152 0.00569 0.0554 -0.01084 0.01147 fixed
sibling_count5 -0.01083 0.005956 -1.819 -0.0225 0.0008413 fixed
sibling_count5+ -0.005929 0.005111 -1.16 -0.01595 0.004088 fixed
birth_order_nonlinear2 0.006418 0.003984 1.611 -0.001391 0.01423 fixed
birth_order_nonlinear3 0.003461 0.004631 0.7474 -0.005616 0.01254 fixed
birth_order_nonlinear4 0.001524 0.005203 0.2928 -0.008674 0.01172 fixed
birth_order_nonlinear5 -0.002664 0.005841 -0.456 -0.01411 0.008785 fixed
birth_order_nonlinear5+ -0.0001667 0.004776 -0.03491 -0.009528 0.009194 fixed
sd_(Intercept).mother_pidlink 0.02114 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1299 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.008777 0.02906 0.302 -0.04818 0.06574 fixed
poly(age, 3, raw = TRUE)1 0.0003881 0.002593 0.1496 -0.004695 0.005471 fixed
poly(age, 3, raw = TRUE)2 0.000006036 0.00007257 0.08317 -0.0001362 0.0001483 fixed
poly(age, 3, raw = TRUE)3 -0.0000001841 0.0000006421 -0.2867 -0.000001443 0.000001074 fixed
male -0.0001087 0.002705 -0.04017 -0.00541 0.005193 fixed
count_birth_order2/2 0.009253 0.00795 1.164 -0.006328 0.02483 fixed
count_birth_order1/3 -0.002485 0.007543 -0.3295 -0.01727 0.0123 fixed
count_birth_order2/3 -0.003798 0.008393 -0.4525 -0.02025 0.01265 fixed
count_birth_order3/3 -0.003063 0.009215 -0.3324 -0.02113 0.015 fixed
count_birth_order1/4 -0.006877 0.008305 -0.8281 -0.02315 0.0094 fixed
count_birth_order2/4 0.02091 0.008908 2.347 0.003452 0.03837 fixed
count_birth_order3/4 0.001834 0.009395 0.1952 -0.01658 0.02025 fixed
count_birth_order4/4 0.002255 0.01012 0.2229 -0.01757 0.02208 fixed
count_birth_order1/5 -0.005294 0.009369 -0.5651 -0.02366 0.01307 fixed
count_birth_order2/5 -0.01186 0.009948 -1.192 -0.03135 0.007642 fixed
count_birth_order3/5 -0.01047 0.01047 -1 -0.03098 0.01005 fixed
count_birth_order4/5 0.008094 0.01105 0.7322 -0.01357 0.02976 fixed
count_birth_order5/5 -0.01984 0.01107 -1.793 -0.04154 0.001853 fixed
count_birth_order1/5+ -0.003349 0.007249 -0.462 -0.01756 0.01086 fixed
count_birth_order2/5+ -0.0008831 0.007537 -0.1172 -0.01566 0.01389 fixed
count_birth_order3/5+ 0.001765 0.007461 0.2366 -0.01286 0.01639 fixed
count_birth_order4/5+ -0.00785 0.007308 -1.074 -0.02217 0.006472 fixed
count_birth_order5/5+ -0.005377 0.007388 -0.7278 -0.01986 0.009103 fixed
count_birth_order5+/5+ -0.005087 0.005909 -0.861 -0.01667 0.006494 fixed
sd_(Intercept).mother_pidlink 0.02122 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1299 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -11872 -11793 5947 -11894 NA NA NA
12 -11870 -11783 5947 -11894 0.007754 1 0.9298
16 -11866 -11751 5949 -11898 4.007 4 0.405
26 -11858 -11671 5955 -11910 12.21 10 0.271

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.06892 0.08321 -0.8282 -0.232 0.09418 fixed
poly(age, 3, raw = TRUE)1 0.009208 0.008926 1.032 -0.008286 0.0267 fixed
poly(age, 3, raw = TRUE)2 -0.0003094 0.0003058 -1.012 -0.0009087 0.0002899 fixed
poly(age, 3, raw = TRUE)3 0.000003375 0.000003352 1.007 -0.000003194 0.000009944 fixed
male -0.003225 0.004207 -0.7666 -0.01147 0.00502 fixed
sibling_count3 -0.004733 0.006873 -0.6886 -0.0182 0.008738 fixed
sibling_count4 0.0002426 0.007083 0.03425 -0.01364 0.01413 fixed
sibling_count5 -0.003574 0.007875 -0.4539 -0.01901 0.01186 fixed
sibling_count5+ 0.000262 0.006798 0.03854 -0.01306 0.01359 fixed
sd_(Intercept).mother_pidlink 0.02884 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1224 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.0693 0.08324 -0.8326 -0.2324 0.09384 fixed
birth_order -0.0003813 0.001296 -0.2942 -0.002922 0.002159 fixed
poly(age, 3, raw = TRUE)1 0.009293 0.008932 1.04 -0.008213 0.0268 fixed
poly(age, 3, raw = TRUE)2 -0.0003114 0.0003059 -1.018 -0.000911 0.0002881 fixed
poly(age, 3, raw = TRUE)3 0.000003381 0.000003352 1.009 -0.000003189 0.000009951 fixed
male -0.00321 0.004208 -0.7627 -0.01146 0.005038 fixed
sibling_count3 -0.004544 0.006904 -0.6581 -0.01808 0.008988 fixed
sibling_count4 0.0006594 0.007225 0.09126 -0.0135 0.01482 fixed
sibling_count5 -0.002889 0.008215 -0.3516 -0.01899 0.01321 fixed
sibling_count5+ 0.001623 0.008225 0.1974 -0.0145 0.01774 fixed
sd_(Intercept).mother_pidlink 0.02889 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1224 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.07098 0.08333 -0.8518 -0.2343 0.09234 fixed
poly(age, 3, raw = TRUE)1 0.00929 0.008936 1.04 -0.008224 0.02681 fixed
poly(age, 3, raw = TRUE)2 -0.0003122 0.0003061 -1.02 -0.0009121 0.0002877 fixed
poly(age, 3, raw = TRUE)3 0.000003395 0.000003355 1.012 -0.00000318 0.000009971 fixed
male -0.003165 0.00421 -0.7517 -0.01142 0.005087 fixed
sibling_count3 -0.00357 0.007046 -0.5067 -0.01738 0.01024 fixed
sibling_count4 0.002033 0.007502 0.271 -0.01267 0.01674 fixed
sibling_count5 -0.001211 0.00862 -0.1405 -0.01811 0.01568 fixed
sibling_count5+ 0.003208 0.008457 0.3793 -0.01337 0.01978 fixed
birth_order_nonlinear2 0.004487 0.005605 0.8006 -0.006498 0.01547 fixed
birth_order_nonlinear3 -0.004523 0.006574 -0.688 -0.01741 0.008361 fixed
birth_order_nonlinear4 -0.001963 0.00802 -0.2447 -0.01768 0.01376 fixed
birth_order_nonlinear5 -0.001521 0.009835 -0.1547 -0.0208 0.01776 fixed
birth_order_nonlinear5+ -0.002346 0.009645 -0.2432 -0.02125 0.01656 fixed
sd_(Intercept).mother_pidlink 0.02914 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1224 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.07149 0.08369 -0.8543 -0.2355 0.09253 fixed
poly(age, 3, raw = TRUE)1 0.009321 0.008976 1.038 -0.008272 0.02691 fixed
poly(age, 3, raw = TRUE)2 -0.000314 0.0003075 -1.021 -0.0009167 0.0002888 fixed
poly(age, 3, raw = TRUE)3 0.000003412 0.000003372 1.012 -0.000003197 0.00001002 fixed
male -0.002905 0.004218 -0.6887 -0.01117 0.005363 fixed
count_birth_order2/2 0.005885 0.01094 0.5379 -0.01556 0.02733 fixed
count_birth_order1/3 -0.004307 0.009118 -0.4724 -0.02218 0.01356 fixed
count_birth_order2/3 0.001187 0.01014 0.1171 -0.01868 0.02106 fixed
count_birth_order3/3 -0.005007 0.01105 -0.4533 -0.02666 0.01664 fixed
count_birth_order1/4 0.0067 0.01046 0.6405 -0.0138 0.0272 fixed
count_birth_order2/4 0.005103 0.01109 0.4601 -0.01664 0.02684 fixed
count_birth_order3/4 0.001785 0.0115 0.1551 -0.02076 0.02433 fixed
count_birth_order4/4 -0.008181 0.01232 -0.6643 -0.03232 0.01596 fixed
count_birth_order1/5 -0.005837 0.01378 -0.4236 -0.03285 0.02117 fixed
count_birth_order2/5 0.01789 0.01487 1.203 -0.01126 0.04703 fixed
count_birth_order3/5 -0.01558 0.01414 -1.102 -0.04328 0.01213 fixed
count_birth_order4/5 -0.006235 0.01379 -0.4522 -0.03326 0.02079 fixed
count_birth_order5/5 0.005751 0.01451 0.3963 -0.02269 0.0342 fixed
count_birth_order1/5+ 0.005701 0.01305 0.4368 -0.01988 0.03128 fixed
count_birth_order2/5+ 0.0003942 0.01316 0.02996 -0.02539 0.02618 fixed
count_birth_order3/5+ -0.001544 0.01284 -0.1203 -0.02671 0.02362 fixed
count_birth_order4/5+ 0.01352 0.01257 1.076 -0.01112 0.03816 fixed
count_birth_order5/5+ -0.002532 0.01201 -0.2108 -0.02608 0.02101 fixed
count_birth_order5+/5+ 0.001374 0.009344 0.147 -0.01694 0.01969 fixed
sd_(Intercept).mother_pidlink 0.02937 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1224 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -4799 -4731 2410 -4821 NA NA NA
12 -4797 -4723 2411 -4821 0.08637 1 0.7688
16 -4791 -4692 2411 -4823 1.815 4 0.7697
26 -4777 -4615 2414 -4829 5.762 10 0.8349

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.07294 0.08327 -0.8759 -0.2362 0.09027 fixed
poly(age, 3, raw = TRUE)1 0.009914 0.00894 1.109 -0.007609 0.02744 fixed
poly(age, 3, raw = TRUE)2 -0.0003348 0.0003063 -1.093 -0.0009353 0.0002656 fixed
poly(age, 3, raw = TRUE)3 0.000003659 0.000003359 1.089 -0.000002924 0.00001024 fixed
male -0.003755 0.004214 -0.8909 -0.01201 0.004505 fixed
sibling_count3 -0.005675 0.007576 -0.7491 -0.02052 0.009173 fixed
sibling_count4 -0.005706 0.007633 -0.7476 -0.02067 0.009254 fixed
sibling_count5 -0.00168 0.008027 -0.2093 -0.01741 0.01405 fixed
sibling_count5+ -0.00157 0.007039 -0.2231 -0.01537 0.01223 fixed
sd_(Intercept).mother_pidlink 0.02896 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.123 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.07311 0.08329 -0.8777 -0.2364 0.09014 fixed
birth_order -0.0002085 0.001148 -0.1815 -0.002459 0.002042 fixed
poly(age, 3, raw = TRUE)1 0.009956 0.008945 1.113 -0.007576 0.02749 fixed
poly(age, 3, raw = TRUE)2 -0.0003358 0.0003064 -1.096 -0.0009364 0.0002648 fixed
poly(age, 3, raw = TRUE)3 0.00000366 0.000003359 1.09 -0.000002924 0.00001024 fixed
male -0.003746 0.004215 -0.8887 -0.01201 0.004516 fixed
sibling_count3 -0.005571 0.007599 -0.7331 -0.02046 0.009323 fixed
sibling_count4 -0.005497 0.007721 -0.7119 -0.02063 0.009636 fixed
sibling_count5 -0.001332 0.008257 -0.1613 -0.01751 0.01485 fixed
sibling_count5+ -0.0008368 0.008118 -0.1031 -0.01675 0.01507 fixed
sd_(Intercept).mother_pidlink 0.02901 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1231 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.07406 0.08335 -0.8886 -0.2374 0.0893 fixed
poly(age, 3, raw = TRUE)1 0.009711 0.008946 1.086 -0.007823 0.02725 fixed
poly(age, 3, raw = TRUE)2 -0.0003301 0.0003065 -1.077 -0.0009309 0.0002706 fixed
poly(age, 3, raw = TRUE)3 0.000003623 0.000003361 1.078 -0.000002963 0.00001021 fixed
male -0.003724 0.004215 -0.8835 -0.01199 0.004537 fixed
sibling_count3 -0.005213 0.00773 -0.6743 -0.02036 0.009938 fixed
sibling_count4 -0.005691 0.007954 -0.7155 -0.02128 0.009899 fixed
sibling_count5 -0.0003466 0.008631 -0.04016 -0.01726 0.01657 fixed
sibling_count5+ 0.0007911 0.008365 0.09457 -0.0156 0.01719 fixed
birth_order_nonlinear2 0.011 0.005694 1.933 -0.000156 0.02217 fixed
birth_order_nonlinear3 -0.001224 0.006636 -0.1845 -0.01423 0.01178 fixed
birth_order_nonlinear4 0.005638 0.007857 0.7176 -0.009762 0.02104 fixed
birth_order_nonlinear5 -0.003972 0.009571 -0.415 -0.02273 0.01479 fixed
birth_order_nonlinear5+ -0.00001819 0.008764 -0.002076 -0.0172 0.01716 fixed
sd_(Intercept).mother_pidlink 0.0295 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1229 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.06463 0.08354 -0.7736 -0.2284 0.09911 fixed
poly(age, 3, raw = TRUE)1 0.008684 0.008969 0.9682 -0.008895 0.02626 fixed
poly(age, 3, raw = TRUE)2 -0.0002926 0.0003074 -0.9519 -0.000895 0.0003099 fixed
poly(age, 3, raw = TRUE)3 0.000003186 0.000003371 0.9452 -0.000003421 0.000009793 fixed
male -0.003631 0.004221 -0.8603 -0.0119 0.004642 fixed
count_birth_order2/2 0.009313 0.01203 0.7738 -0.01427 0.0329 fixed
count_birth_order1/3 -0.01236 0.01012 -1.221 -0.03219 0.00748 fixed
count_birth_order2/3 0.0116 0.01102 1.052 -0.01001 0.03321 fixed
count_birth_order3/3 -0.003213 0.01222 -0.2628 -0.02717 0.02074 fixed
count_birth_order1/4 -0.01075 0.01098 -0.9789 -0.03228 0.01078 fixed
count_birth_order2/4 0.01045 0.01158 0.9025 -0.01224 0.03314 fixed
count_birth_order3/4 -0.002847 0.01253 -0.2271 -0.02741 0.02172 fixed
count_birth_order4/4 -0.007088 0.01361 -0.5207 -0.03377 0.01959 fixed
count_birth_order1/5 0.01516 0.01328 1.141 -0.01087 0.04118 fixed
count_birth_order2/5 0.008984 0.01362 0.6598 -0.01771 0.03567 fixed
count_birth_order3/5 -0.01683 0.01401 -1.201 -0.0443 0.01063 fixed
count_birth_order4/5 -0.007029 0.01427 -0.4927 -0.03499 0.02093 fixed
count_birth_order5/5 0.003691 0.01449 0.2548 -0.0247 0.03208 fixed
count_birth_order1/5+ 0.004812 0.01198 0.4017 -0.01866 0.02829 fixed
count_birth_order2/5+ -0.003156 0.01266 -0.2493 -0.02797 0.02166 fixed
count_birth_order3/5+ 0.0007469 0.0122 0.06123 -0.02316 0.02465 fixed
count_birth_order4/5+ 0.01669 0.01185 1.408 -0.006541 0.03992 fixed
count_birth_order5/5+ -0.009497 0.01255 -0.7565 -0.0341 0.01511 fixed
count_birth_order5+/5+ 0.000137 0.009443 0.01451 -0.01837 0.01865 fixed
sd_(Intercept).mother_pidlink 0.03021 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1227 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -4793 -4725 2408 -4815 NA NA NA
12 -4791 -4717 2408 -4815 0.03277 1 0.8563
16 -4789 -4689 2410 -4821 5.725 4 0.2207
26 -4781 -4619 2416 -4833 11.8 10 0.2989

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.0498 0.0826 -0.603 -0.2117 0.1121 fixed
poly(age, 3, raw = TRUE)1 0.007002 0.008874 0.789 -0.01039 0.0244 fixed
poly(age, 3, raw = TRUE)2 -0.0002367 0.0003043 -0.778 -0.0008331 0.0003596 fixed
poly(age, 3, raw = TRUE)3 0.000002649 0.000003339 0.7933 -0.000003896 0.000009194 fixed
male -0.002382 0.004174 -0.5707 -0.01056 0.005798 fixed
sibling_count3 -0.00575 0.00661 -0.8699 -0.01871 0.007206 fixed
sibling_count4 0.002621 0.006909 0.3793 -0.01092 0.01616 fixed
sibling_count5 -0.001736 0.007888 -0.22 -0.0172 0.01373 fixed
sibling_count5+ -0.001253 0.006699 -0.187 -0.01438 0.01188 fixed
sd_(Intercept).mother_pidlink 0.02772 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1205 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.04939 0.08263 -0.5978 -0.2113 0.1126 fixed
birth_order 0.0003003 0.001312 0.2289 -0.002272 0.002872 fixed
poly(age, 3, raw = TRUE)1 0.006922 0.008882 0.7793 -0.01049 0.02433 fixed
poly(age, 3, raw = TRUE)2 -0.0002347 0.0003045 -0.7709 -0.0008314 0.000362 fixed
poly(age, 3, raw = TRUE)3 0.00000264 0.00000334 0.7903 -0.000003907 0.000009186 fixed
male -0.002386 0.004174 -0.5716 -0.01057 0.005795 fixed
sibling_count3 -0.005903 0.006644 -0.8883 -0.01893 0.00712 fixed
sibling_count4 0.002296 0.007054 0.3255 -0.01153 0.01612 fixed
sibling_count5 -0.002258 0.008213 -0.2749 -0.01835 0.01384 fixed
sibling_count5+ -0.002324 0.008174 -0.2844 -0.01835 0.0137 fixed
sd_(Intercept).mother_pidlink 0.02771 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1205 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.05159 0.08271 -0.6238 -0.2137 0.1105 fixed
poly(age, 3, raw = TRUE)1 0.007037 0.008885 0.792 -0.01038 0.02445 fixed
poly(age, 3, raw = TRUE)2 -0.0002395 0.0003046 -0.7862 -0.0008365 0.0003575 fixed
poly(age, 3, raw = TRUE)3 0.0000027 0.000003342 0.8077 -0.000003851 0.00000925 fixed
male -0.002368 0.004176 -0.567 -0.01055 0.005817 fixed
sibling_count3 -0.00439 0.006788 -0.6468 -0.01769 0.008914 fixed
sibling_count4 0.00465 0.007327 0.6346 -0.009711 0.01901 fixed
sibling_count5 0.00009162 0.008584 0.01067 -0.01673 0.01692 fixed
sibling_count5+ -0.0008097 0.008416 -0.09621 -0.0173 0.01568 fixed
birth_order_nonlinear2 0.005265 0.005486 0.9597 -0.005488 0.01602 fixed
birth_order_nonlinear3 -0.005105 0.006471 -0.7888 -0.01779 0.007579 fixed
birth_order_nonlinear4 -0.002657 0.008121 -0.3272 -0.01857 0.01326 fixed
birth_order_nonlinear5 0.002499 0.01001 0.2497 -0.01712 0.02212 fixed
birth_order_nonlinear5+ 0.003725 0.009781 0.3809 -0.01544 0.0229 fixed
sd_(Intercept).mother_pidlink 0.02781 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1205 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.05244 0.08302 -0.6316 -0.2152 0.1103 fixed
poly(age, 3, raw = TRUE)1 0.007219 0.008922 0.8091 -0.01027 0.0247 fixed
poly(age, 3, raw = TRUE)2 -0.0002477 0.000306 -0.8097 -0.0008474 0.0003519 fixed
poly(age, 3, raw = TRUE)3 0.000002801 0.000003358 0.8341 -0.000003781 0.000009382 fixed
male -0.002176 0.004184 -0.5202 -0.01038 0.006024 fixed
count_birth_order2/2 0.004609 0.01046 0.4405 -0.0159 0.02511 fixed
count_birth_order1/3 -0.004567 0.008804 -0.5187 -0.02182 0.01269 fixed
count_birth_order2/3 -0.002892 0.009793 -0.2953 -0.02209 0.0163 fixed
count_birth_order3/3 -0.00515 0.01055 -0.4882 -0.02583 0.01553 fixed
count_birth_order1/4 0.007362 0.01033 0.7125 -0.01289 0.02761 fixed
count_birth_order2/4 0.01055 0.01086 0.972 -0.01073 0.03184 fixed
count_birth_order3/4 0.00207 0.01127 0.1836 -0.02002 0.02416 fixed
count_birth_order4/4 -0.007332 0.01224 -0.5988 -0.03133 0.01667 fixed
count_birth_order1/5 -0.006338 0.01351 -0.4691 -0.03282 0.02014 fixed
count_birth_order2/5 0.02056 0.01509 1.362 -0.009021 0.05014 fixed
count_birth_order3/5 -0.01572 0.01472 -1.068 -0.04457 0.01313 fixed
count_birth_order4/5 -0.004808 0.01417 -0.3393 -0.03258 0.02296 fixed
count_birth_order5/5 0.009342 0.01505 0.6207 -0.02016 0.03884 fixed
count_birth_order1/5+ -0.001178 0.01317 -0.0895 -0.02698 0.02463 fixed
count_birth_order2/5+ 0.001312 0.01332 0.0985 -0.02479 0.02741 fixed
count_birth_order3/5+ -0.009649 0.01272 -0.7588 -0.03457 0.01527 fixed
count_birth_order4/5+ 0.007954 0.01277 0.6227 -0.01708 0.03299 fixed
count_birth_order5/5+ -0.002177 0.01199 -0.1815 -0.02568 0.02133 fixed
count_birth_order5+/5+ 0.002798 0.009292 0.3011 -0.01541 0.02101 fixed
sd_(Intercept).mother_pidlink 0.02796 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1206 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -4839 -4771 2430 -4861 NA NA NA
12 -4837 -4762 2430 -4861 0.05301 1 0.8179
16 -4832 -4732 2432 -4864 2.872 4 0.5795
26 -4817 -4656 2434 -4869 5.404 10 0.8626

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Sector_Finance, insurance, real estate and business services

birthorder <- birthorder %>% mutate(outcome = `Sector_Finance, insurance, real estate and business services`)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.07441 0.09233 -0.8059 -0.2554 0.1066 fixed
poly(age, 3, raw = TRUE)1 0.02539 0.008265 3.072 0.009189 0.04159 fixed
poly(age, 3, raw = TRUE)2 -0.0006471 0.0002316 -2.795 -0.001101 -0.0001933 fixed
poly(age, 3, raw = TRUE)3 0.000005081 0.000002052 2.476 0.000001059 0.000009104 fixed
male 0.0594 0.008556 6.943 0.04264 0.07617 fixed
sibling_count3 -0.007552 0.01779 -0.4245 -0.04241 0.02731 fixed
sibling_count4 -0.02384 0.01795 -1.328 -0.05903 0.01135 fixed
sibling_count5 -0.02626 0.01865 -1.407 -0.06282 0.01031 fixed
sibling_count5+ -0.05315 0.0145 -3.664 -0.08158 -0.02472 fixed
sd_(Intercept).mother_pidlink 0.1381 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3954 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.07441 0.09233 -0.8059 -0.2554 0.1066 fixed
birth_order 0.0000267 0.001729 0.01544 -0.003363 0.003416 fixed
poly(age, 3, raw = TRUE)1 0.02538 0.008279 3.066 0.009153 0.04161 fixed
poly(age, 3, raw = TRUE)2 -0.0006469 0.0002322 -2.785 -0.001102 -0.0001917 fixed
poly(age, 3, raw = TRUE)3 0.000005079 0.000002058 2.467 0.000001045 0.000009114 fixed
male 0.0594 0.008556 6.943 0.04263 0.07617 fixed
sibling_count3 -0.007559 0.01779 -0.4248 -0.04243 0.02732 fixed
sibling_count4 -0.02386 0.01799 -1.327 -0.05911 0.01139 fixed
sibling_count5 -0.02628 0.01875 -1.402 -0.06303 0.01046 fixed
sibling_count5+ -0.05324 0.01573 -3.386 -0.08406 -0.02242 fixed
sd_(Intercept).mother_pidlink 0.1381 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3954 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.07779 0.09248 -0.8411 -0.2591 0.1035 fixed
poly(age, 3, raw = TRUE)1 0.02597 0.008287 3.134 0.00973 0.04221 fixed
poly(age, 3, raw = TRUE)2 -0.0006675 0.0002324 -2.872 -0.001123 -0.000212 fixed
poly(age, 3, raw = TRUE)3 0.000005269 0.000002059 2.558 0.000001233 0.000009306 fixed
male 0.05941 0.008557 6.943 0.04264 0.07618 fixed
sibling_count3 -0.01047 0.01806 -0.58 -0.04586 0.02492 fixed
sibling_count4 -0.02378 0.01848 -1.287 -0.06 0.01243 fixed
sibling_count5 -0.02538 0.01939 -1.309 -0.06338 0.01262 fixed
sibling_count5+ -0.0482 0.01647 -2.927 -0.08049 -0.01592 fixed
birth_order_nonlinear2 -0.0008122 0.01249 -0.06505 -0.02528 0.02366 fixed
birth_order_nonlinear3 0.01115 0.0145 0.7694 -0.01726 0.03957 fixed
birth_order_nonlinear4 -0.01657 0.0163 -1.016 -0.04852 0.01538 fixed
birth_order_nonlinear5 -0.005199 0.01831 -0.2839 -0.04109 0.03069 fixed
birth_order_nonlinear5+ -0.01029 0.01518 -0.6782 -0.04004 0.01945 fixed
sd_(Intercept).mother_pidlink 0.1378 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3955 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.06935 0.09283 -0.7471 -0.2513 0.1126 fixed
poly(age, 3, raw = TRUE)1 0.02651 0.008296 3.195 0.01025 0.04277 fixed
poly(age, 3, raw = TRUE)2 -0.0006742 0.0002325 -2.899 -0.00113 -0.0002185 fixed
poly(age, 3, raw = TRUE)3 0.000005249 0.00000206 2.548 0.000001211 0.000009288 fixed
male 0.05988 0.008562 6.994 0.0431 0.07666 fixed
count_birth_order2/2 -0.04629 0.02487 -1.861 -0.09504 0.002463 fixed
count_birth_order1/3 -0.01612 0.02398 -0.6721 -0.06313 0.03089 fixed
count_birth_order2/3 -0.02574 0.02667 -0.9651 -0.07801 0.02653 fixed
count_birth_order3/3 -0.04184 0.02927 -1.429 -0.09922 0.01553 fixed
count_birth_order1/4 -0.05395 0.02639 -2.044 -0.1057 -0.00223 fixed
count_birth_order2/4 -0.02768 0.0283 -0.978 -0.08314 0.02779 fixed
count_birth_order3/4 -0.02782 0.02984 -0.9325 -0.0863 0.03066 fixed
count_birth_order4/4 -0.05826 0.03211 -1.814 -0.1212 0.004687 fixed
count_birth_order1/5 -0.05281 0.02975 -1.775 -0.1111 0.00549 fixed
count_birth_order2/5 -0.02327 0.03158 -0.7369 -0.08516 0.03862 fixed
count_birth_order3/5 -0.02847 0.03321 -0.8573 -0.09357 0.03662 fixed
count_birth_order4/5 -0.09112 0.03505 -2.6 -0.1598 -0.02242 fixed
count_birth_order5/5 -0.03226 0.03512 -0.9184 -0.1011 0.03658 fixed
count_birth_order1/5+ -0.08272 0.023 -3.596 -0.1278 -0.03763 fixed
count_birth_order2/5+ -0.06344 0.02391 -2.654 -0.1103 -0.01658 fixed
count_birth_order3/5+ -0.04458 0.02366 -1.884 -0.09096 0.001796 fixed
count_birth_order4/5+ -0.07335 0.02319 -3.163 -0.1188 -0.0279 fixed
count_birth_order5/5+ -0.0759 0.02343 -3.239 -0.1218 -0.02997 fixed
count_birth_order5+/5+ -0.07606 0.01898 -4.007 -0.1133 -0.03886 fixed
sd_(Intercept).mother_pidlink 0.1378 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3956 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 10644 10723 -5311 10622 NA NA NA
12 10646 10732 -5311 10622 0.0002311 1 0.9879
16 10651 10766 -5309 10619 3.022 4 0.5542
26 10661 10848 -5305 10609 9.616 10 0.4748

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5687 0.2886 -1.971 -1.134 -0.003089 fixed
poly(age, 3, raw = TRUE)1 0.0821 0.03095 2.653 0.02144 0.1428 fixed
poly(age, 3, raw = TRUE)2 -0.002522 0.00106 -2.378 -0.0046 -0.0004438 fixed
poly(age, 3, raw = TRUE)3 0.00002507 0.00001162 2.157 0.000002289 0.00004785 fixed
male 0.05418 0.01459 3.715 0.0256 0.08277 fixed
sibling_count3 -0.03401 0.02387 -1.424 -0.0808 0.01278 fixed
sibling_count4 -0.04521 0.02461 -1.837 -0.09345 0.00303 fixed
sibling_count5 -0.09774 0.02738 -3.57 -0.1514 -0.04408 fixed
sibling_count5+ -0.08594 0.02364 -3.636 -0.1323 -0.03961 fixed
sd_(Intercept).mother_pidlink 0.1082 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4226 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5687 0.2886 -1.97 -1.134 -0.002995 fixed
birth_order -0.00005679 0.004499 -0.01262 -0.008875 0.008761 fixed
poly(age, 3, raw = TRUE)1 0.08212 0.03097 2.651 0.02141 0.1428 fixed
poly(age, 3, raw = TRUE)2 -0.002522 0.001061 -2.378 -0.004602 -0.0004432 fixed
poly(age, 3, raw = TRUE)3 0.00002507 0.00001163 2.157 0.000002286 0.00004786 fixed
male 0.05418 0.01459 3.714 0.02559 0.08278 fixed
sibling_count3 -0.03398 0.02398 -1.417 -0.08098 0.01303 fixed
sibling_count4 -0.04515 0.02511 -1.798 -0.09435 0.00406 fixed
sibling_count5 -0.09764 0.02856 -3.419 -0.1536 -0.04166 fixed
sibling_count5+ -0.08574 0.02859 -2.999 -0.1418 -0.0297 fixed
sd_(Intercept).mother_pidlink 0.1083 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4226 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5737 0.2889 -1.986 -1.14 -0.007463 fixed
poly(age, 3, raw = TRUE)1 0.08249 0.03098 2.662 0.02176 0.1432 fixed
poly(age, 3, raw = TRUE)2 -0.002533 0.001061 -2.387 -0.004613 -0.0004533 fixed
poly(age, 3, raw = TRUE)3 0.00002515 0.00001163 2.162 0.000002352 0.00004795 fixed
male 0.05423 0.01459 3.715 0.02562 0.08283 fixed
sibling_count3 -0.02672 0.02446 -1.092 -0.07466 0.02122 fixed
sibling_count4 -0.03303 0.02605 -1.268 -0.08408 0.01803 fixed
sibling_count5 -0.08304 0.02994 -2.773 -0.1417 -0.02435 fixed
sibling_count5+ -0.07479 0.02938 -2.546 -0.1324 -0.01721 fixed
birth_order_nonlinear2 0.004422 0.01941 0.2278 -0.03362 0.04246 fixed
birth_order_nonlinear3 -0.03044 0.02277 -1.337 -0.07506 0.01418 fixed
birth_order_nonlinear4 -0.02453 0.02778 -0.8831 -0.07898 0.02992 fixed
birth_order_nonlinear5 -0.01353 0.03407 -0.3972 -0.0803 0.05324 fixed
birth_order_nonlinear5+ -0.002097 0.03345 -0.06269 -0.06766 0.06347 fixed
sd_(Intercept).mother_pidlink 0.1081 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4227 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5355 0.2901 -1.846 -1.104 0.03302 fixed
poly(age, 3, raw = TRUE)1 0.07959 0.03111 2.558 0.01861 0.1406 fixed
poly(age, 3, raw = TRUE)2 -0.002426 0.001066 -2.276 -0.004515 -0.0003364 fixed
poly(age, 3, raw = TRUE)3 0.00002391 0.00001169 2.046 0.000001001 0.00004681 fixed
male 0.05452 0.01462 3.73 0.02587 0.08318 fixed
count_birth_order2/2 -0.03903 0.03789 -1.03 -0.1133 0.03524 fixed
count_birth_order1/3 -0.04971 0.03161 -1.573 -0.1117 0.01223 fixed
count_birth_order2/3 -0.01491 0.03514 -0.4244 -0.08378 0.05396 fixed
count_birth_order3/3 -0.08276 0.03828 -2.162 -0.1578 -0.007725 fixed
count_birth_order1/4 -0.05126 0.03626 -1.414 -0.1223 0.01981 fixed
count_birth_order2/4 -0.05105 0.03844 -1.328 -0.1264 0.0243 fixed
count_birth_order3/4 -0.09079 0.03987 -2.277 -0.1689 -0.01265 fixed
count_birth_order4/4 -0.03882 0.04268 -0.9094 -0.1225 0.04484 fixed
count_birth_order1/5 -0.1067 0.04777 -2.234 -0.2003 -0.0131 fixed
count_birth_order2/5 -0.08707 0.05153 -1.69 -0.1881 0.01394 fixed
count_birth_order3/5 -0.1337 0.04899 -2.728 -0.2297 -0.03764 fixed
count_birth_order4/5 -0.1599 0.04779 -3.345 -0.2535 -0.06621 fixed
count_birth_order5/5 -0.05661 0.0503 -1.126 -0.1552 0.04196 fixed
count_birth_order1/5+ -0.09304 0.04523 -2.057 -0.1817 -0.004385 fixed
count_birth_order2/5+ -0.07479 0.0456 -1.64 -0.1642 0.01458 fixed
count_birth_order3/5+ -0.08138 0.0445 -1.829 -0.1686 0.005839 fixed
count_birth_order4/5+ -0.1182 0.04357 -2.712 -0.2035 -0.03277 fixed
count_birth_order5/5+ -0.1354 0.04163 -3.253 -0.217 -0.05382 fixed
count_birth_order5+/5+ -0.0913 0.03241 -2.817 -0.1548 -0.02779 fixed
sd_(Intercept).mother_pidlink 0.1074 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.423 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 4337 4406 -2158 4315 NA NA NA
12 4339 4414 -2158 4315 0.00009047 1 0.9924
16 4344 4444 -2156 4312 2.997 4 0.5583
26 4357 4518 -2152 4305 7.716 10 0.6566

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5662 0.2875 -1.97 -1.13 -0.002767 fixed
poly(age, 3, raw = TRUE)1 0.08105 0.03086 2.626 0.02057 0.1415 fixed
poly(age, 3, raw = TRUE)2 -0.002488 0.001058 -2.353 -0.004561 -0.0004157 fixed
poly(age, 3, raw = TRUE)3 0.00002469 0.0000116 2.129 0.000001961 0.00004741 fixed
male 0.05379 0.01454 3.699 0.02529 0.0823 fixed
sibling_count3 -0.01975 0.0262 -0.7539 -0.0711 0.0316 fixed
sibling_count4 -0.03216 0.0264 -1.218 -0.08391 0.01959 fixed
sibling_count5 -0.07215 0.02778 -2.597 -0.1266 -0.0177 fixed
sibling_count5+ -0.07094 0.02436 -2.912 -0.1187 -0.0232 fixed
sd_(Intercept).mother_pidlink 0.1094 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4225 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5664 0.2875 -1.97 -1.13 -0.002844 fixed
birth_order -0.0002095 0.00397 -0.05278 -0.007991 0.007572 fixed
poly(age, 3, raw = TRUE)1 0.0811 0.03088 2.626 0.02058 0.1416 fixed
poly(age, 3, raw = TRUE)2 -0.002489 0.001058 -2.353 -0.004563 -0.0004162 fixed
poly(age, 3, raw = TRUE)3 0.00002469 0.0000116 2.129 0.00000196 0.00004742 fixed
male 0.0538 0.01455 3.699 0.02529 0.08231 fixed
sibling_count3 -0.01965 0.02628 -0.7476 -0.07115 0.03186 fixed
sibling_count4 -0.03195 0.02671 -1.196 -0.08429 0.0204 fixed
sibling_count5 -0.0718 0.02857 -2.513 -0.1278 -0.01579 fixed
sibling_count5+ -0.0702 0.02809 -2.499 -0.1253 -0.01515 fixed
sd_(Intercept).mother_pidlink 0.1096 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4225 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.578 0.2878 -2.008 -1.142 -0.01393 fixed
poly(age, 3, raw = TRUE)1 0.08201 0.03089 2.655 0.02147 0.1426 fixed
poly(age, 3, raw = TRUE)2 -0.002519 0.001058 -2.38 -0.004593 -0.0004445 fixed
poly(age, 3, raw = TRUE)3 0.00002491 0.0000116 2.147 0.00000217 0.00004766 fixed
male 0.05398 0.01455 3.71 0.02546 0.08249 fixed
sibling_count3 -0.01423 0.02673 -0.5325 -0.06663 0.03816 fixed
sibling_count4 -0.01948 0.02751 -0.7081 -0.07341 0.03444 fixed
sibling_count5 -0.05724 0.02986 -1.917 -0.1158 0.00128 fixed
sibling_count5+ -0.05202 0.02894 -1.798 -0.1087 0.004701 fixed
birth_order_nonlinear2 0.01072 0.01963 0.5459 -0.02776 0.04919 fixed
birth_order_nonlinear3 -0.02222 0.02288 -0.9713 -0.06707 0.02262 fixed
birth_order_nonlinear4 -0.03943 0.0271 -1.455 -0.09254 0.01368 fixed
birth_order_nonlinear5 -0.007367 0.03301 -0.2232 -0.07206 0.05733 fixed
birth_order_nonlinear5+ -0.01766 0.03028 -0.5834 -0.077 0.04168 fixed
sd_(Intercept).mother_pidlink 0.1108 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4222 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.512 0.2882 -1.776 -1.077 0.05292 fixed
poly(age, 3, raw = TRUE)1 0.07721 0.03094 2.495 0.01656 0.1379 fixed
poly(age, 3, raw = TRUE)2 -0.002343 0.001061 -2.209 -0.004421 -0.0002639 fixed
poly(age, 3, raw = TRUE)3 0.00002292 0.00001163 1.97 0.0000001211 0.00004571 fixed
male 0.05323 0.01456 3.656 0.02469 0.08177 fixed
count_birth_order2/2 -0.06461 0.04148 -1.558 -0.1459 0.01669 fixed
count_birth_order1/3 -0.0366 0.03492 -1.048 -0.105 0.03184 fixed
count_birth_order2/3 -0.01324 0.03804 -0.348 -0.08779 0.06131 fixed
count_birth_order3/3 -0.08953 0.04217 -2.123 -0.1722 -0.006883 fixed
count_birth_order1/4 -0.08289 0.0379 -2.187 -0.1572 -0.008614 fixed
count_birth_order2/4 0.00127 0.03994 0.03181 -0.077 0.07955 fixed
count_birth_order3/4 -0.08315 0.04323 -1.923 -0.1679 0.001588 fixed
count_birth_order4/4 -0.04866 0.04696 -1.036 -0.1407 0.04338 fixed
count_birth_order1/5 -0.06946 0.04581 -1.516 -0.1592 0.02033 fixed
count_birth_order2/5 -0.1132 0.04698 -2.409 -0.2052 -0.0211 fixed
count_birth_order3/5 -0.1311 0.04834 -2.712 -0.2258 -0.03636 fixed
count_birth_order4/5 -0.1211 0.04921 -2.46 -0.2175 -0.02459 fixed
count_birth_order5/5 -0.03203 0.04997 -0.641 -0.13 0.06591 fixed
count_birth_order1/5+ -0.08991 0.04133 -2.176 -0.1709 -0.008917 fixed
count_birth_order2/5+ -0.04979 0.04368 -1.14 -0.1354 0.03581 fixed
count_birth_order3/5+ -0.04301 0.04208 -1.022 -0.1255 0.03947 fixed
count_birth_order4/5+ -0.1407 0.04088 -3.442 -0.2208 -0.06057 fixed
count_birth_order5/5+ -0.1239 0.0433 -2.86 -0.2087 -0.03899 fixed
count_birth_order5+/5+ -0.09545 0.0326 -2.928 -0.1593 -0.03156 fixed
sd_(Intercept).mother_pidlink 0.1099 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.422 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 4369 4437 -2173 4347 NA NA NA
12 4371 4445 -2173 4347 0.00247 1 0.9604
16 4375 4474 -2171 4343 4.144 4 0.3868
26 4378 4539 -2163 4326 16.88 10 0.07713

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5595 0.2914 -1.92 -1.131 0.01169 fixed
poly(age, 3, raw = TRUE)1 0.0817 0.03131 2.609 0.02033 0.1431 fixed
poly(age, 3, raw = TRUE)2 -0.0025 0.001074 -2.329 -0.004605 -0.0003961 fixed
poly(age, 3, raw = TRUE)3 0.00002474 0.00001178 2.1 0.000001645 0.00004783 fixed
male 0.05654 0.01472 3.841 0.02769 0.08539 fixed
sibling_count3 -0.05676 0.02338 -2.427 -0.1026 -0.01093 fixed
sibling_count4 -0.04937 0.02446 -2.019 -0.0973 -0.001435 fixed
sibling_count5 -0.1102 0.02796 -3.94 -0.1649 -0.05536 fixed
sibling_count5+ -0.09375 0.02374 -3.949 -0.1403 -0.04722 fixed
sd_(Intercept).mother_pidlink 0.1108 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.422 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5626 0.2915 -1.93 -1.134 0.008759 fixed
birth_order -0.002165 0.004637 -0.467 -0.01125 0.006922 fixed
poly(age, 3, raw = TRUE)1 0.08229 0.03134 2.626 0.02087 0.1437 fixed
poly(age, 3, raw = TRUE)2 -0.002516 0.001074 -2.342 -0.004621 -0.0004102 fixed
poly(age, 3, raw = TRUE)3 0.00002481 0.00001178 2.105 0.000001714 0.00004791 fixed
male 0.05657 0.01472 3.843 0.02772 0.08543 fixed
sibling_count3 -0.05566 0.02351 -2.368 -0.1017 -0.009586 fixed
sibling_count4 -0.04701 0.02497 -1.882 -0.09596 0.001936 fixed
sibling_count5 -0.1064 0.02911 -3.654 -0.1634 -0.04932 fixed
sibling_count5+ -0.08602 0.02896 -2.97 -0.1428 -0.02925 fixed
sd_(Intercept).mother_pidlink 0.1114 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4219 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5677 0.2918 -1.946 -1.14 0.004187 fixed
poly(age, 3, raw = TRUE)1 0.08282 0.03135 2.642 0.02137 0.1443 fixed
poly(age, 3, raw = TRUE)2 -0.002528 0.001075 -2.352 -0.004634 -0.0004216 fixed
poly(age, 3, raw = TRUE)3 0.00002487 0.00001179 2.108 0.000001751 0.00004798 fixed
male 0.05634 0.01473 3.826 0.02748 0.08521 fixed
sibling_count3 -0.04861 0.024 -2.025 -0.09565 -0.001557 fixed
sibling_count4 -0.03533 0.02592 -1.363 -0.08613 0.01548 fixed
sibling_count5 -0.09458 0.0304 -3.111 -0.1542 -0.035 fixed
sibling_count5+ -0.07499 0.0298 -2.517 -0.1334 -0.01659 fixed
birth_order_nonlinear2 -0.01061 0.01931 -0.5491 -0.04846 0.02725 fixed
birth_order_nonlinear3 -0.03414 0.02279 -1.499 -0.0788 0.01051 fixed
birth_order_nonlinear4 -0.03378 0.0286 -1.181 -0.08984 0.02228 fixed
birth_order_nonlinear5 -0.01735 0.03526 -0.4922 -0.08645 0.05175 fixed
birth_order_nonlinear5+ -0.0251 0.03452 -0.727 -0.09277 0.04257 fixed
sd_(Intercept).mother_pidlink 0.111 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4221 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.5353 0.2929 -1.827 -1.109 0.03882 fixed
poly(age, 3, raw = TRUE)1 0.08019 0.03148 2.547 0.01849 0.1419 fixed
poly(age, 3, raw = TRUE)2 -0.002433 0.00108 -2.254 -0.004549 -0.0003172 fixed
poly(age, 3, raw = TRUE)3 0.00002378 0.00001185 2.007 0.0000005629 0.00004701 fixed
male 0.0564 0.01476 3.822 0.02748 0.08533 fixed
count_birth_order2/2 -0.04001 0.03684 -1.086 -0.1122 0.03219 fixed
count_birth_order1/3 -0.06363 0.03106 -2.049 -0.1245 -0.002752 fixed
count_birth_order2/3 -0.05555 0.03455 -1.608 -0.1233 0.01216 fixed
count_birth_order3/3 -0.09984 0.03721 -2.683 -0.1728 -0.0269 fixed
count_birth_order1/4 -0.04157 0.03645 -1.141 -0.113 0.02987 fixed
count_birth_order2/4 -0.06433 0.0383 -1.68 -0.1394 0.01074 fixed
count_birth_order3/4 -0.09248 0.03975 -2.326 -0.1704 -0.01457 fixed
count_birth_order4/4 -0.05644 0.04319 -1.307 -0.1411 0.02821 fixed
count_birth_order1/5 -0.127 0.04766 -2.664 -0.2204 -0.03354 fixed
count_birth_order2/5 -0.1031 0.05323 -1.937 -0.2074 0.00122 fixed
count_birth_order3/5 -0.1444 0.05191 -2.782 -0.2462 -0.04268 fixed
count_birth_order4/5 -0.1675 0.04997 -3.352 -0.2654 -0.06954 fixed
count_birth_order5/5 -0.06378 0.05308 -1.202 -0.1678 0.04025 fixed
count_birth_order1/5+ -0.08858 0.04645 -1.907 -0.1796 0.002461 fixed
count_birth_order2/5+ -0.08499 0.04697 -1.809 -0.1771 0.007076 fixed
count_birth_order3/5+ -0.08575 0.04485 -1.912 -0.1737 0.002157 fixed
count_birth_order4/5+ -0.1208 0.04505 -2.683 -0.2091 -0.03255 fixed
count_birth_order5/5+ -0.1342 0.04229 -3.172 -0.217 -0.05126 fixed
count_birth_order5+/5+ -0.11 0.03283 -3.35 -0.1743 -0.04564 fixed
sd_(Intercept).mother_pidlink 0.1107 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4224 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 4263 4331 -2120 4241 NA NA NA
12 4264 4339 -2120 4240 0.2157 1 0.6424
16 4270 4369 -2119 4238 2.603 4 0.6264
26 4284 4445 -2116 4232 5.553 10 0.8513

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Sector_Manufacturing

birthorder <- birthorder %>% mutate(outcome = Sector_Manufacturing)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.5117 0.09457 5.41 0.3263 0.6971 fixed
poly(age, 3, raw = TRUE)1 -0.01479 0.008466 -1.747 -0.03138 0.001801 fixed
poly(age, 3, raw = TRUE)2 0.0002146 0.0002372 0.9045 -0.0002504 0.0006795 fixed
poly(age, 3, raw = TRUE)3 -0.0000006725 0.000002103 -0.3198 -0.000004794 0.000003449 fixed
male -0.02311 0.008759 -2.639 -0.04028 -0.005944 fixed
sibling_count3 -0.006571 0.01824 -0.3603 -0.04232 0.02917 fixed
sibling_count4 -0.00687 0.01841 -0.3732 -0.04295 0.02921 fixed
sibling_count5 0.01292 0.01913 0.6752 -0.02458 0.05042 fixed
sibling_count5+ 0.01186 0.01487 0.7972 -0.01729 0.04101 fixed
sd_(Intercept).mother_pidlink 0.1442 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4041 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.5117 0.09458 5.41 0.3263 0.6971 fixed
birth_order -0.0001019 0.001772 -0.05749 -0.003576 0.003372 fixed
poly(age, 3, raw = TRUE)1 -0.01476 0.008481 -1.741 -0.03139 0.001859 fixed
poly(age, 3, raw = TRUE)2 0.0002136 0.0002379 0.8976 -0.0002528 0.0006799 fixed
poly(age, 3, raw = TRUE)3 -0.0000006633 0.000002109 -0.3145 -0.000004797 0.00000347 fixed
male -0.02311 0.00876 -2.638 -0.04028 -0.005941 fixed
sibling_count3 -0.006544 0.01824 -0.3587 -0.0423 0.02921 fixed
sibling_count4 -0.006808 0.01844 -0.3691 -0.04296 0.02934 fixed
sibling_count5 0.01303 0.01923 0.6776 -0.02465 0.05071 fixed
sibling_count5+ 0.01221 0.01612 0.7577 -0.01938 0.04381 fixed
sd_(Intercept).mother_pidlink 0.1442 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4041 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.5096 0.09474 5.379 0.3239 0.6953 fixed
poly(age, 3, raw = TRUE)1 -0.01469 0.008489 -1.731 -0.03133 0.001947 fixed
poly(age, 3, raw = TRUE)2 0.0002074 0.0002381 0.8713 -0.0002592 0.000674 fixed
poly(age, 3, raw = TRUE)3 -0.0000005849 0.00000211 -0.2772 -0.000004721 0.000003551 fixed
male -0.02303 0.008761 -2.629 -0.04021 -0.005862 fixed
sibling_count3 -0.005231 0.01851 -0.2826 -0.04152 0.03105 fixed
sibling_count4 -0.007222 0.01895 -0.3811 -0.04436 0.02992 fixed
sibling_count5 0.01484 0.01988 0.7462 -0.02413 0.0538 fixed
sibling_count5+ 0.01629 0.01688 0.965 -0.0168 0.04938 fixed
birth_order_nonlinear2 0.008384 0.01278 0.6562 -0.01666 0.03343 fixed
birth_order_nonlinear3 -0.003908 0.01484 -0.2634 -0.03299 0.02517 fixed
birth_order_nonlinear4 0.01104 0.01668 0.6616 -0.02166 0.04373 fixed
birth_order_nonlinear5 -0.01119 0.01874 -0.5973 -0.04792 0.02554 fixed
birth_order_nonlinear5+ -0.004781 0.01554 -0.3076 -0.03525 0.02568 fixed
sd_(Intercept).mother_pidlink 0.1441 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4042 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.4939 0.09509 5.194 0.3075 0.6803 fixed
poly(age, 3, raw = TRUE)1 -0.01421 0.008498 -1.673 -0.03087 0.002443 fixed
poly(age, 3, raw = TRUE)2 0.0001951 0.0002382 0.8191 -0.0002718 0.0006621 fixed
poly(age, 3, raw = TRUE)3 -0.000000494 0.000002111 -0.234 -0.000004632 0.000003644 fixed
male -0.02312 0.008766 -2.638 -0.0403 -0.005942 fixed
count_birth_order2/2 0.03648 0.02545 1.433 -0.01341 0.08636 fixed
count_birth_order1/3 0.01974 0.02457 0.8034 -0.02841 0.06788 fixed
count_birth_order2/3 0.009603 0.02731 0.3516 -0.04393 0.06314 fixed
count_birth_order3/3 -0.02071 0.02998 -0.6908 -0.07947 0.03805 fixed
count_birth_order1/4 0.008555 0.02702 0.3166 -0.04441 0.06152 fixed
count_birth_order2/4 0.0266 0.02898 0.9179 -0.0302 0.08341 fixed
count_birth_order3/4 -0.03689 0.03056 -1.207 -0.09678 0.02299 fixed
count_birth_order4/4 0.02723 0.03289 0.8279 -0.03723 0.09169 fixed
count_birth_order1/5 0.04031 0.03046 1.323 -0.01939 0.1 fixed
count_birth_order2/5 0.02362 0.03234 0.7304 -0.03976 0.087 fixed
count_birth_order3/5 0.03578 0.03401 1.052 -0.03088 0.1024 fixed
count_birth_order4/5 0.008249 0.03589 0.2298 -0.0621 0.0786 fixed
count_birth_order5/5 0.01433 0.03597 0.3985 -0.05616 0.08483 fixed
count_birth_order1/5+ 0.01689 0.02356 0.717 -0.02928 0.06306 fixed
count_birth_order2/5+ 0.01899 0.02448 0.7757 -0.02899 0.06698 fixed
count_birth_order3/5+ 0.04577 0.02423 1.889 -0.001725 0.09326 fixed
count_birth_order4/5+ 0.04063 0.02374 1.711 -0.005906 0.08717 fixed
count_birth_order5/5+ 0.01524 0.024 0.6352 -0.03179 0.06228 fixed
count_birth_order5+/5+ 0.02177 0.01945 1.119 -0.01635 0.05989 fixed
sd_(Intercept).mother_pidlink 0.1442 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4041 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 11106 11186 -5542 11084 NA NA NA
12 11108 11195 -5542 11084 0.003397 1 0.9535
16 11114 11229 -5541 11082 2.159 4 0.7066
26 11124 11311 -5536 11072 9.91 10 0.4484

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.6945 0.2861 2.428 0.1338 1.255 fixed
poly(age, 3, raw = TRUE)1 -0.03818 0.03069 -1.244 -0.09833 0.02196 fixed
poly(age, 3, raw = TRUE)2 0.001072 0.001051 1.02 -0.000988 0.003133 fixed
poly(age, 3, raw = TRUE)3 -0.00001043 0.00001152 -0.9051 -0.00003302 0.00001216 fixed
male -0.01913 0.01445 -1.324 -0.04746 0.00919 fixed
sibling_count3 0.001268 0.02378 0.05334 -0.04534 0.04787 fixed
sibling_count4 0.02087 0.02454 0.8503 -0.02723 0.06896 fixed
sibling_count5 0.02786 0.02734 1.019 -0.02573 0.08144 fixed
sibling_count5+ 0.01755 0.02361 0.7432 -0.02873 0.06382 fixed
sd_(Intercept).mother_pidlink 0.1269 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4136 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.6924 0.2862 2.42 0.1316 1.253 fixed
birth_order -0.001776 0.004471 -0.3974 -0.01054 0.006986 fixed
poly(age, 3, raw = TRUE)1 -0.03775 0.03071 -1.229 -0.09793 0.02244 fixed
poly(age, 3, raw = TRUE)2 0.001062 0.001052 1.009 -0.0009998 0.003123 fixed
poly(age, 3, raw = TRUE)3 -0.00001039 0.00001153 -0.9014 -0.00003298 0.0000122 fixed
male -0.01907 0.01445 -1.32 -0.0474 0.009257 fixed
sibling_count3 0.002154 0.02388 0.09018 -0.04466 0.04896 fixed
sibling_count4 0.02283 0.02503 0.9121 -0.02623 0.07188 fixed
sibling_count5 0.03107 0.02851 1.09 -0.02481 0.08695 fixed
sibling_count5+ 0.02391 0.02854 0.838 -0.03202 0.07985 fixed
sd_(Intercept).mother_pidlink 0.1266 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4138 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.7097 0.2864 2.478 0.1484 1.271 fixed
poly(age, 3, raw = TRUE)1 -0.04004 0.03071 -1.304 -0.1002 0.02015 fixed
poly(age, 3, raw = TRUE)2 0.001135 0.001052 1.079 -0.0009269 0.003197 fixed
poly(age, 3, raw = TRUE)3 -0.00001111 0.00001153 -0.9635 -0.00003371 0.00001149 fixed
male -0.01949 0.01446 -1.348 -0.04783 0.00884 fixed
sibling_count3 0.002157 0.02435 0.08859 -0.04557 0.04989 fixed
sibling_count4 0.01402 0.02595 0.5401 -0.03685 0.06489 fixed
sibling_count5 0.02613 0.02986 0.875 -0.0324 0.08466 fixed
sibling_count5+ 0.02047 0.02931 0.6984 -0.03697 0.07791 fixed
birth_order_nonlinear2 0.008278 0.01916 0.4321 -0.02927 0.04583 fixed
birth_order_nonlinear3 -0.003105 0.02248 -0.1381 -0.04717 0.04096 fixed
birth_order_nonlinear4 0.0427 0.02744 1.556 -0.01109 0.09649 fixed
birth_order_nonlinear5 -0.029 0.03366 -0.8614 -0.09497 0.03698 fixed
birth_order_nonlinear5+ -0.005666 0.03318 -0.1707 -0.07071 0.05938 fixed
sd_(Intercept).mother_pidlink 0.128 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4133 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.7378 0.2875 2.566 0.1742 1.301 fixed
poly(age, 3, raw = TRUE)1 -0.04289 0.03084 -1.391 -0.1033 0.01755 fixed
poly(age, 3, raw = TRUE)2 0.001232 0.001057 1.166 -0.0008394 0.003303 fixed
poly(age, 3, raw = TRUE)3 -0.00001216 0.00001159 -1.05 -0.00003487 0.00001055 fixed
male -0.01952 0.01448 -1.348 -0.0479 0.008861 fixed
count_birth_order2/2 0.004085 0.0374 0.1092 -0.06922 0.07739 fixed
count_birth_order1/3 0.01393 0.03133 0.4445 -0.04748 0.07534 fixed
count_birth_order2/3 0.007459 0.03483 0.2142 -0.0608 0.07572 fixed
count_birth_order3/3 -0.02574 0.03793 -0.6785 -0.1001 0.04861 fixed
count_birth_order1/4 -0.01898 0.03594 -0.528 -0.08942 0.05146 fixed
count_birth_order2/4 0.0258 0.0381 0.6772 -0.04887 0.1005 fixed
count_birth_order3/4 0.04904 0.0395 1.241 -0.02838 0.1265 fixed
count_birth_order4/4 0.05322 0.04228 1.259 -0.02965 0.1361 fixed
count_birth_order1/5 0.01815 0.04733 0.3835 -0.07462 0.1109 fixed
count_birth_order2/5 0.08403 0.05105 1.646 -0.01603 0.1841 fixed
count_birth_order3/5 -0.02305 0.04853 -0.4749 -0.1182 0.07206 fixed
count_birth_order4/5 0.07164 0.04733 1.513 -0.02114 0.1644 fixed
count_birth_order5/5 -0.001846 0.04981 -0.03705 -0.09947 0.09578 fixed
count_birth_order1/5+ 0.03997 0.04483 0.8917 -0.04789 0.1278 fixed
count_birth_order2/5+ -0.008876 0.04517 -0.1965 -0.09742 0.07966 fixed
count_birth_order3/5+ 0.03244 0.04408 0.7359 -0.05396 0.1188 fixed
count_birth_order4/5+ 0.06151 0.04315 1.426 -0.02305 0.1461 fixed
count_birth_order5/5+ -0.01123 0.04124 -0.2723 -0.09205 0.06959 fixed
count_birth_order5+/5+ 0.01346 0.0322 0.4178 -0.04966 0.07658 fixed
sd_(Intercept).mother_pidlink 0.1285 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4133 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 4272 4341 -2125 4250 NA NA NA
12 4274 4349 -2125 4250 0.16 1 0.6892
16 4277 4377 -2123 4245 4.804 4 0.3081
26 4290 4451 -2119 4238 7.584 10 0.6694

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.6938 0.2847 2.437 0.1359 1.252 fixed
poly(age, 3, raw = TRUE)1 -0.03908 0.03056 -1.279 -0.09899 0.02082 fixed
poly(age, 3, raw = TRUE)2 0.001103 0.001047 1.053 -0.00095 0.003155 fixed
poly(age, 3, raw = TRUE)3 -0.00001075 0.00001148 -0.936 -0.00003326 0.00001176 fixed
male -0.01987 0.01439 -1.381 -0.04808 0.008339 fixed
sibling_count3 0.02057 0.02606 0.7896 -0.0305 0.07164 fixed
sibling_count4 0.01883 0.02628 0.7167 -0.03267 0.07034 fixed
sibling_count5 0.02997 0.02767 1.083 -0.02427 0.08421 fixed
sibling_count5+ 0.02952 0.02426 1.217 -0.01803 0.07707 fixed
sd_(Intercept).mother_pidlink 0.1274 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4132 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.6907 0.2847 2.426 0.1327 1.249 fixed
birth_order -0.003004 0.003942 -0.7619 -0.01073 0.004723 fixed
poly(age, 3, raw = TRUE)1 -0.0384 0.03058 -1.256 -0.09833 0.02153 fixed
poly(age, 3, raw = TRUE)2 0.001086 0.001048 1.037 -0.0009671 0.003139 fixed
poly(age, 3, raw = TRUE)3 -0.0000107 0.00001148 -0.9316 -0.00003321 0.00001181 fixed
male -0.01978 0.0144 -1.374 -0.04799 0.00844 fixed
sibling_count3 0.02208 0.02613 0.8452 -0.02913 0.07329 fixed
sibling_count4 0.02188 0.02657 0.8235 -0.0302 0.07397 fixed
sibling_count5 0.03503 0.02845 1.231 -0.02074 0.0908 fixed
sibling_count5+ 0.04012 0.02797 1.435 -0.01469 0.09494 fixed
sd_(Intercept).mother_pidlink 0.1267 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4134 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.7126 0.2849 2.501 0.1543 1.271 fixed
poly(age, 3, raw = TRUE)1 -0.04117 0.03058 -1.347 -0.1011 0.01876 fixed
poly(age, 3, raw = TRUE)2 0.001175 0.001048 1.122 -0.000878 0.003229 fixed
poly(age, 3, raw = TRUE)3 -0.00001154 0.00001149 -1.005 -0.00003406 0.00001097 fixed
male -0.02032 0.01439 -1.412 -0.04853 0.007889 fixed
sibling_count3 0.02206 0.02657 0.83 -0.03003 0.07414 fixed
sibling_count4 0.0125 0.02736 0.4569 -0.04113 0.06613 fixed
sibling_count5 0.03038 0.02971 1.023 -0.02785 0.08862 fixed
sibling_count5+ 0.03026 0.0288 1.051 -0.02619 0.0867 fixed
birth_order_nonlinear2 0.000648 0.01936 0.03347 -0.0373 0.03859 fixed
birth_order_nonlinear3 -0.006387 0.02258 -0.2829 -0.05064 0.03786 fixed
birth_order_nonlinear4 0.04502 0.02675 1.683 -0.007412 0.09744 fixed
birth_order_nonlinear5 -0.04231 0.03258 -1.299 -0.1062 0.02155 fixed
birth_order_nonlinear5+ -0.003707 0.03001 -0.1235 -0.06252 0.05511 fixed
sd_(Intercept).mother_pidlink 0.1294 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4125 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.715 0.2859 2.501 0.1548 1.275 fixed
poly(age, 3, raw = TRUE)1 -0.04177 0.03069 -1.361 -0.1019 0.01837 fixed
poly(age, 3, raw = TRUE)2 0.001195 0.001052 1.137 -0.0008661 0.003257 fixed
poly(age, 3, raw = TRUE)3 -0.00001175 0.00001154 -1.019 -0.00003436 0.00001086 fixed
male -0.02002 0.01443 -1.387 -0.0483 0.008265 fixed
count_birth_order2/2 0.0098 0.041 0.239 -0.07055 0.09015 fixed
count_birth_order1/3 0.03993 0.03463 1.153 -0.02795 0.1078 fixed
count_birth_order2/3 0.02923 0.03772 0.7751 -0.04469 0.1032 fixed
count_birth_order3/3 -0.01499 0.0418 -0.3586 -0.09692 0.06694 fixed
count_birth_order1/4 0.008483 0.03758 0.2257 -0.06518 0.08214 fixed
count_birth_order2/4 0.00627 0.0396 0.1583 -0.07134 0.08388 fixed
count_birth_order3/4 0.02997 0.04285 0.6994 -0.05402 0.114 fixed
count_birth_order4/4 0.06428 0.04654 1.381 -0.02694 0.1555 fixed
count_birth_order1/5 0.008951 0.04542 0.1971 -0.08008 0.09798 fixed
count_birth_order2/5 0.06832 0.04657 1.467 -0.02296 0.1596 fixed
count_birth_order3/5 0.02295 0.04791 0.4791 -0.07095 0.1169 fixed
count_birth_order4/5 0.07555 0.04877 1.549 -0.02005 0.1711 fixed
count_birth_order5/5 -0.01083 0.04952 -0.2187 -0.1079 0.08622 fixed
count_birth_order1/5+ 0.03972 0.04098 0.9691 -0.04061 0.12 fixed
count_birth_order2/5+ 0.00776 0.04329 0.1793 -0.07709 0.09261 fixed
count_birth_order3/5+ 0.04269 0.04171 1.024 -0.03905 0.1244 fixed
count_birth_order4/5+ 0.07782 0.04052 1.921 -0.001596 0.1572 fixed
count_birth_order5/5+ -0.007665 0.04291 -0.1786 -0.09176 0.07643 fixed
count_birth_order5+/5+ 0.02949 0.03239 0.9104 -0.034 0.09298 fixed
sd_(Intercept).mother_pidlink 0.1294 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4129 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 4295 4364 -2137 4273 NA NA NA
12 4297 4371 -2136 4273 0.585 1 0.4444
16 4298 4398 -2133 4266 6.406 4 0.1708
26 4314 4476 -2131 4262 4.039 10 0.9455

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.6847 0.2888 2.371 0.1186 1.251 fixed
poly(age, 3, raw = TRUE)1 -0.03708 0.03103 -1.195 -0.0979 0.02374 fixed
poly(age, 3, raw = TRUE)2 0.001033 0.001064 0.9707 -0.001053 0.003118 fixed
poly(age, 3, raw = TRUE)3 -0.000009884 0.00001168 -0.8464 -0.00003277 0.000013 fixed
male -0.02402 0.01458 -1.648 -0.0526 0.004552 fixed
sibling_count3 0.006789 0.02327 0.2917 -0.03882 0.0524 fixed
sibling_count4 0.01743 0.02436 0.7156 -0.03032 0.06518 fixed
sibling_count5 0.03358 0.0279 1.204 -0.0211 0.08826 fixed
sibling_count5+ 0.0147 0.02369 0.6205 -0.03174 0.06114 fixed
sd_(Intercept).mother_pidlink 0.1271 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4135 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.6811 0.2889 2.357 0.1148 1.247 fixed
birth_order -0.002423 0.004604 -0.5263 -0.01145 0.0066 fixed
poly(age, 3, raw = TRUE)1 -0.03639 0.03106 -1.172 -0.09727 0.02448 fixed
poly(age, 3, raw = TRUE)2 0.001015 0.001065 0.9533 -0.001072 0.003102 fixed
poly(age, 3, raw = TRUE)3 -0.000009793 0.00001168 -0.8384 -0.00003269 0.0000131 fixed
male -0.024 0.01458 -1.646 -0.05258 0.004575 fixed
sibling_count3 0.008022 0.02339 0.343 -0.03782 0.05386 fixed
sibling_count4 0.02008 0.02487 0.8073 -0.02867 0.06882 fixed
sibling_count5 0.03781 0.02903 1.302 -0.0191 0.09471 fixed
sibling_count5+ 0.02339 0.02887 0.81 -0.0332 0.07997 fixed
sd_(Intercept).mother_pidlink 0.1268 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4136 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.6914 0.2892 2.391 0.1246 1.258 fixed
poly(age, 3, raw = TRUE)1 -0.03822 0.03107 -1.23 -0.09911 0.02267 fixed
poly(age, 3, raw = TRUE)2 0.001074 0.001065 1.008 -0.001014 0.003161 fixed
poly(age, 3, raw = TRUE)3 -0.00001037 0.00001169 -0.8876 -0.00003328 0.00001253 fixed
male -0.02434 0.01459 -1.669 -0.05293 0.004249 fixed
sibling_count3 0.009206 0.02388 0.3855 -0.03759 0.056 fixed
sibling_count4 0.01429 0.0258 0.5538 -0.03628 0.06485 fixed
sibling_count5 0.03383 0.03029 1.117 -0.02555 0.0932 fixed
sibling_count5+ 0.02141 0.02969 0.7211 -0.03678 0.0796 fixed
birth_order_nonlinear2 0.01259 0.01907 0.6599 -0.0248 0.04997 fixed
birth_order_nonlinear3 -0.008404 0.02251 -0.3733 -0.05252 0.03572 fixed
birth_order_nonlinear4 0.03288 0.02827 1.163 -0.02252 0.08828 fixed
birth_order_nonlinear5 -0.01862 0.03484 -0.5346 -0.08691 0.04966 fixed
birth_order_nonlinear5+ -0.01146 0.03423 -0.3349 -0.07856 0.05563 fixed
sd_(Intercept).mother_pidlink 0.127 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4136 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.7099 0.2904 2.445 0.1407 1.279 fixed
poly(age, 3, raw = TRUE)1 -0.04044 0.03121 -1.296 -0.1016 0.02073 fixed
poly(age, 3, raw = TRUE)2 0.001151 0.00107 1.076 -0.0009463 0.003249 fixed
poly(age, 3, raw = TRUE)3 -0.00001124 0.00001175 -0.9567 -0.00003426 0.00001178 fixed
male -0.02448 0.01462 -1.675 -0.05314 0.00417 fixed
count_birth_order2/2 0.01734 0.03639 0.4765 -0.05399 0.08867 fixed
count_birth_order1/3 0.01849 0.03079 0.6004 -0.04186 0.07884 fixed
count_birth_order2/3 0.02155 0.03424 0.6293 -0.04556 0.08866 fixed
count_birth_order3/3 -0.009869 0.03687 -0.2676 -0.08214 0.0624 fixed
count_birth_order1/4 -0.01049 0.03613 -0.2903 -0.08131 0.06033 fixed
count_birth_order2/4 0.027 0.03796 0.7113 -0.0474 0.1014 fixed
count_birth_order3/4 0.03907 0.03939 0.992 -0.03813 0.1163 fixed
count_birth_order4/4 0.05482 0.04279 1.281 -0.02905 0.1387 fixed
count_birth_order1/5 0.03448 0.04724 0.73 -0.0581 0.1271 fixed
count_birth_order2/5 0.06907 0.05274 1.31 -0.03429 0.1724 fixed
count_birth_order3/5 0.004936 0.05143 0.09599 -0.09586 0.1057 fixed
count_birth_order4/5 0.06178 0.0495 1.248 -0.03524 0.1588 fixed
count_birth_order5/5 0.02736 0.05258 0.5203 -0.0757 0.1304 fixed
count_birth_order1/5+ 0.05413 0.04604 1.176 -0.03611 0.1444 fixed
count_birth_order2/5+ 0.02108 0.04654 0.4529 -0.07015 0.1123 fixed
count_birth_order3/5+ 0.006787 0.04444 0.1527 -0.08031 0.09388 fixed
count_birth_order4/5+ 0.05472 0.04463 1.226 -0.03274 0.1422 fixed
count_birth_order5/5+ -0.00152 0.0419 -0.03627 -0.08363 0.08059 fixed
count_birth_order5+/5+ 0.01146 0.03262 0.3515 -0.05247 0.0754 fixed
sd_(Intercept).mother_pidlink 0.1273 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.4139 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 4196 4264 -2087 4174 NA NA NA
12 4198 4272 -2087 4174 0.2797 1 0.5969
16 4203 4302 -2085 4171 3.113 4 0.5391
26 4219 4380 -2084 4167 3.496 10 0.9672

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Sector_Mining and quarrying

birthorder <- birthorder %>% mutate(outcome = `Sector_Mining and quarrying`)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.03035 0.03991 -0.7605 -0.1086 0.04787 fixed
poly(age, 3, raw = TRUE)1 0.002784 0.003572 0.7794 -0.004217 0.009786 fixed
poly(age, 3, raw = TRUE)2 -0.00004645 0.0001001 -0.4641 -0.0002426 0.0001497 fixed
poly(age, 3, raw = TRUE)3 0.0000002049 0.000000887 0.231 -0.000001534 0.000001943 fixed
male 0.02971 0.0037 8.03 0.02246 0.03696 fixed
sibling_count3 -0.0008367 0.007684 -0.1089 -0.0159 0.01422 fixed
sibling_count4 0.004985 0.007755 0.6429 -0.01021 0.02018 fixed
sibling_count5 0.006666 0.008057 0.8274 -0.009125 0.02246 fixed
sibling_count5+ 0.001122 0.006266 0.179 -0.01116 0.0134 fixed
sd_(Intercept).mother_pidlink 0.05891 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1712 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.03038 0.03991 -0.7613 -0.1086 0.04784 fixed
birth_order 0.0005056 0.0007473 0.6765 -0.0009592 0.00197 fixed
poly(age, 3, raw = TRUE)1 0.002644 0.003579 0.7388 -0.00437 0.009658 fixed
poly(age, 3, raw = TRUE)2 -0.00004135 0.0001004 -0.412 -0.0002381 0.0001554 fixed
poly(age, 3, raw = TRUE)3 0.0000001591 0.0000008896 0.1788 -0.000001585 0.000001903 fixed
male 0.0297 0.0037 8.027 0.02245 0.03695 fixed
sibling_count3 -0.0009676 0.007687 -0.1259 -0.01603 0.0141 fixed
sibling_count4 0.004674 0.007769 0.6016 -0.01055 0.0199 fixed
sibling_count5 0.006127 0.008097 0.7568 -0.009742 0.022 fixed
sibling_count5+ -0.000655 0.006794 -0.09641 -0.01397 0.01266 fixed
sd_(Intercept).mother_pidlink 0.05893 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1712 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.02744 0.03998 -0.6864 -0.1058 0.05091 fixed
poly(age, 3, raw = TRUE)1 0.002552 0.003582 0.7125 -0.004468 0.009573 fixed
poly(age, 3, raw = TRUE)2 -0.00003768 0.0001004 -0.3752 -0.0002345 0.0001592 fixed
poly(age, 3, raw = TRUE)3 0.0000001159 0.0000008902 0.1301 -0.000001629 0.000001861 fixed
male 0.02973 0.0037 8.035 0.02248 0.03698 fixed
sibling_count3 0.0002275 0.007802 0.02916 -0.01506 0.01552 fixed
sibling_count4 0.004558 0.007985 0.5709 -0.01109 0.02021 fixed
sibling_count5 0.00631 0.008377 0.7532 -0.01011 0.02273 fixed
sibling_count5+ -0.0008504 0.007119 -0.1195 -0.0148 0.0131 fixed
birth_order_nonlinear2 -0.004173 0.0054 -0.7729 -0.01476 0.00641 fixed
birth_order_nonlinear3 -0.005299 0.00627 -0.8451 -0.01759 0.006991 fixed
birth_order_nonlinear4 0.006777 0.00705 0.9613 -0.007041 0.02059 fixed
birth_order_nonlinear5 -0.00166 0.00792 -0.2096 -0.01718 0.01386 fixed
birth_order_nonlinear5+ 0.002234 0.006562 0.3404 -0.01063 0.0151 fixed
sd_(Intercept).mother_pidlink 0.05905 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1712 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.0343 0.04011 -0.8551 -0.1129 0.04432 fixed
poly(age, 3, raw = TRUE)1 0.002683 0.003585 0.7484 -0.004343 0.009709 fixed
poly(age, 3, raw = TRUE)2 -0.00004222 0.0001005 -0.4202 -0.0002392 0.0001547 fixed
poly(age, 3, raw = TRUE)3 0.0000001617 0.0000008903 0.1816 -0.000001583 0.000001907 fixed
male 0.0297 0.003701 8.025 0.02245 0.03695 fixed
count_birth_order2/2 0.01138 0.01075 1.058 -0.009698 0.03246 fixed
count_birth_order1/3 0.005769 0.01037 0.5566 -0.01455 0.02608 fixed
count_birth_order2/3 0.01716 0.01153 1.489 -0.005428 0.03975 fixed
count_birth_order3/3 -0.01888 0.01265 -1.492 -0.04367 0.005919 fixed
count_birth_order1/4 0.01393 0.0114 1.222 -0.008416 0.03628 fixed
count_birth_order2/4 -0.005393 0.01223 -0.441 -0.02936 0.01858 fixed
count_birth_order3/4 0.009254 0.01289 0.7177 -0.01602 0.03453 fixed
count_birth_order4/4 0.02237 0.01388 1.612 -0.004829 0.04958 fixed
count_birth_order1/5 0.005794 0.01286 0.4507 -0.0194 0.03099 fixed
count_birth_order2/5 0.009803 0.01365 0.7183 -0.01694 0.03655 fixed
count_birth_order3/5 0.007786 0.01435 0.5424 -0.02035 0.03592 fixed
count_birth_order4/5 0.01815 0.01515 1.198 -0.01154 0.04784 fixed
count_birth_order5/5 0.01756 0.01518 1.157 -0.01219 0.04732 fixed
count_birth_order1/5+ 0.01252 0.009942 1.259 -0.006965 0.03201 fixed
count_birth_order2/5+ -0.01148 0.01033 -1.111 -0.03173 0.008772 fixed
count_birth_order3/5+ 0.007394 0.01023 0.723 -0.01265 0.02744 fixed
count_birth_order4/5+ 0.01005 0.01002 1.003 -0.009592 0.02969 fixed
count_birth_order5/5+ 0.001259 0.01013 0.1243 -0.01859 0.02111 fixed
count_birth_order5+/5+ 0.007163 0.0082 0.8736 -0.008909 0.02324 fixed
sd_(Intercept).mother_pidlink 0.05892 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1712 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -5714 -5635 2868 -5736 NA NA NA
12 -5712 -5626 2868 -5736 0.4579 1 0.4986
16 -5707 -5592 2870 -5739 3.207 4 0.5238
26 -5703 -5517 2878 -5755 16.09 10 0.09704

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.0135 0.1052 -0.1283 -0.2197 0.1927 fixed
poly(age, 3, raw = TRUE)1 0.0004965 0.01128 0.044 -0.02162 0.02261 fixed
poly(age, 3, raw = TRUE)2 0.00002584 0.0003865 0.06685 -0.0007317 0.0007834 fixed
poly(age, 3, raw = TRUE)3 -0.0000008291 0.000004237 -0.1957 -0.000009134 0.000007476 fixed
male 0.02643 0.005311 4.977 0.01602 0.03684 fixed
sibling_count3 0.01414 0.008772 1.612 -0.003053 0.03133 fixed
sibling_count4 0.01755 0.00906 1.937 -0.0002053 0.03531 fixed
sibling_count5 0.004679 0.0101 0.4631 -0.01512 0.02448 fixed
sibling_count5+ 0.01454 0.008725 1.666 -0.002565 0.03164 fixed
sd_(Intercept).mother_pidlink 0.05107 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1508 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.01511 0.1052 -0.1436 -0.2213 0.1911 fixed
birth_order -0.00126 0.001647 -0.7652 -0.004487 0.001967 fixed
poly(age, 3, raw = TRUE)1 0.0008181 0.01129 0.07246 -0.02131 0.02295 fixed
poly(age, 3, raw = TRUE)2 0.00001765 0.0003867 0.04563 -0.0007403 0.0007755 fixed
poly(age, 3, raw = TRUE)3 -0.0000007941 0.000004238 -0.1874 -0.0000091 0.000007512 fixed
male 0.02649 0.005312 4.987 0.01608 0.0369 fixed
sibling_count3 0.01478 0.008814 1.676 -0.002499 0.03205 fixed
sibling_count4 0.01895 0.009244 2.05 0.000829 0.03706 fixed
sibling_count5 0.006955 0.01054 0.6599 -0.0137 0.02761 fixed
sibling_count5+ 0.01908 0.01055 1.809 -0.001596 0.03975 fixed
sd_(Intercept).mother_pidlink 0.05147 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1506 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.01618 0.1053 -0.1537 -0.2226 0.1902 fixed
poly(age, 3, raw = TRUE)1 0.00098 0.01129 0.08678 -0.02115 0.02311 fixed
poly(age, 3, raw = TRUE)2 0.0000145 0.0003868 0.03749 -0.0007437 0.0007727 fixed
poly(age, 3, raw = TRUE)3 -0.0000007877 0.00000424 -0.1858 -0.000009099 0.000007523 fixed
male 0.02645 0.005314 4.978 0.01604 0.03687 fixed
sibling_count3 0.01569 0.008983 1.747 -0.001915 0.0333 fixed
sibling_count4 0.02248 0.009577 2.347 0.003705 0.04125 fixed
sibling_count5 0.01117 0.01103 1.013 -0.01044 0.03278 fixed
sibling_count5+ 0.02095 0.01082 1.936 -0.0002616 0.04216 fixed
birth_order_nonlinear2 -0.008145 0.007027 -1.159 -0.02192 0.005629 fixed
birth_order_nonlinear3 -0.006853 0.008248 -0.8308 -0.02302 0.009314 fixed
birth_order_nonlinear4 -0.01932 0.01007 -1.919 -0.03906 0.0004151 fixed
birth_order_nonlinear5 -0.01062 0.01235 -0.8596 -0.03483 0.01359 fixed
birth_order_nonlinear5+ -0.007731 0.01221 -0.6332 -0.03166 0.0162 fixed
sd_(Intercept).mother_pidlink 0.05144 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1507 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.009966 0.1058 -0.09421 -0.2173 0.1974 fixed
poly(age, 3, raw = TRUE)1 0.0004782 0.01135 0.04215 -0.02176 0.02272 fixed
poly(age, 3, raw = TRUE)2 0.00003556 0.0003887 0.09147 -0.0007264 0.0007975 fixed
poly(age, 3, raw = TRUE)3 -0.000001051 0.000004262 -0.2466 -0.000009405 0.000007303 fixed
male 0.0264 0.005325 4.957 0.01596 0.03683 fixed
count_birth_order2/2 -0.01641 0.01373 -1.195 -0.04332 0.0105 fixed
count_birth_order1/3 0.00795 0.01153 0.6896 -0.01465 0.03055 fixed
count_birth_order2/3 0.0165 0.01281 1.288 -0.00861 0.04161 fixed
count_birth_order3/3 0.0005857 0.01395 0.04198 -0.02676 0.02793 fixed
count_birth_order1/4 0.0207 0.01322 1.565 -0.005221 0.04661 fixed
count_birth_order2/4 0.005796 0.01401 0.4136 -0.02167 0.03326 fixed
count_birth_order3/4 0.008942 0.01453 0.6155 -0.01953 0.03742 fixed
count_birth_order4/4 0.01142 0.01555 0.734 -0.01907 0.0419 fixed
count_birth_order1/5 0.007885 0.01741 0.4528 -0.02624 0.04201 fixed
count_birth_order2/5 -0.006315 0.01877 -0.3363 -0.04311 0.03048 fixed
count_birth_order3/5 0.008757 0.01785 0.4907 -0.02622 0.04373 fixed
count_birth_order4/5 -0.01357 0.01741 -0.7797 -0.04769 0.02054 fixed
count_birth_order5/5 -0.0001377 0.01832 -0.007518 -0.03604 0.03576 fixed
count_birth_order1/5+ 0.02285 0.01649 1.385 -0.009476 0.05517 fixed
count_birth_order2/5+ 0.008602 0.01661 0.5177 -0.02396 0.04116 fixed
count_birth_order3/5+ 0.01917 0.01621 1.182 -0.01261 0.05094 fixed
count_birth_order4/5+ -0.01033 0.01587 -0.651 -0.04143 0.02077 fixed
count_birth_order5/5+ 0.006311 0.01516 0.4162 -0.02341 0.03603 fixed
count_birth_order5+/5+ 0.01036 0.01186 0.8735 -0.01289 0.03362 fixed
sd_(Intercept).mother_pidlink 0.05117 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1509 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -3080 -3011 1551 -3102 NA NA NA
12 -3078 -3004 1551 -3102 0.5788 1 0.4468
16 -3074 -2975 1553 -3106 3.483 4 0.4804
26 -3059 -2897 1555 -3111 4.848 10 0.9011

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.01139 0.1044 -0.1092 -0.216 0.1932 fixed
poly(age, 3, raw = TRUE)1 0.0004498 0.0112 0.04014 -0.02151 0.02241 fixed
poly(age, 3, raw = TRUE)2 0.00002628 0.000384 0.06844 -0.0007263 0.0007788 fixed
poly(age, 3, raw = TRUE)3 -0.0000008199 0.00000421 -0.1948 -0.000009072 0.000007432 fixed
male 0.02609 0.005275 4.946 0.01575 0.03643 fixed
sibling_count3 0.008191 0.009579 0.8551 -0.01058 0.02697 fixed
sibling_count4 0.02248 0.009664 2.326 0.003539 0.04142 fixed
sibling_count5 -0.0006205 0.01018 -0.06093 -0.02058 0.01934 fixed
sibling_count5+ 0.01265 0.008926 1.417 -0.00485 0.03014 fixed
sd_(Intercept).mother_pidlink 0.05052 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1503 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.01242 0.1044 -0.119 -0.217 0.1922 fixed
birth_order -0.0009043 0.001448 -0.6243 -0.003743 0.001934 fixed
poly(age, 3, raw = TRUE)1 0.0006657 0.01121 0.05938 -0.02131 0.02264 fixed
poly(age, 3, raw = TRUE)2 0.00002094 0.0003841 0.05452 -0.0007318 0.0007737 fixed
poly(age, 3, raw = TRUE)3 -0.0000008013 0.000004211 -0.1903 -0.000009054 0.000007451 fixed
male 0.02613 0.005276 4.953 0.01579 0.03647 fixed
sibling_count3 0.008654 0.00961 0.9005 -0.01018 0.02749 fixed
sibling_count4 0.0234 0.009779 2.393 0.004236 0.04257 fixed
sibling_count5 0.000899 0.01048 0.08581 -0.01964 0.02143 fixed
sibling_count5+ 0.01586 0.0103 1.54 -0.004323 0.03603 fixed
sd_(Intercept).mother_pidlink 0.05085 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1502 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.01457 0.1045 -0.1394 -0.2194 0.1902 fixed
poly(age, 3, raw = TRUE)1 0.000977 0.01121 0.08712 -0.021 0.02296 fixed
poly(age, 3, raw = TRUE)2 0.00001225 0.0003842 0.03189 -0.0007408 0.0007653 fixed
poly(age, 3, raw = TRUE)3 -0.0000007374 0.000004213 -0.175 -0.000008995 0.00000752 fixed
male 0.0261 0.005277 4.946 0.01575 0.03644 fixed
sibling_count3 0.009587 0.00977 0.9813 -0.009562 0.02874 fixed
sibling_count4 0.02691 0.01006 2.674 0.007189 0.04663 fixed
sibling_count5 0.004979 0.01093 0.4555 -0.01644 0.0264 fixed
sibling_count5+ 0.01859 0.01059 1.755 -0.002169 0.03936 fixed
birth_order_nonlinear2 -0.006724 0.007086 -0.9489 -0.02061 0.007165 fixed
birth_order_nonlinear3 -0.00603 0.008265 -0.7296 -0.02223 0.01017 fixed
birth_order_nonlinear4 -0.01962 0.009795 -2.003 -0.03882 -0.0004223 fixed
birth_order_nonlinear5 -0.008111 0.01193 -0.6798 -0.0315 0.01527 fixed
birth_order_nonlinear5+ -0.007123 0.01101 -0.6469 -0.02871 0.01446 fixed
sd_(Intercept).mother_pidlink 0.05104 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1502 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.01177 0.1048 -0.1123 -0.2172 0.1936 fixed
poly(age, 3, raw = TRUE)1 0.0009535 0.01125 0.08475 -0.0211 0.023 fixed
poly(age, 3, raw = TRUE)2 0.00001652 0.0003856 0.04284 -0.0007392 0.0007723 fixed
poly(age, 3, raw = TRUE)3 -0.000000813 0.000004229 -0.1922 -0.000009102 0.000007476 fixed
male 0.02624 0.005288 4.961 0.01587 0.0366 fixed
count_birth_order2/2 -0.01773 0.01501 -1.182 -0.04714 0.01168 fixed
count_birth_order1/3 -0.0005947 0.0127 -0.04683 -0.02549 0.0243 fixed
count_birth_order2/3 0.01466 0.01383 1.06 -0.01244 0.04177 fixed
count_birth_order3/3 -0.009318 0.01532 -0.608 -0.03935 0.02072 fixed
count_birth_order1/4 0.03084 0.01378 2.238 0.003834 0.05785 fixed
count_birth_order2/4 0.009355 0.01452 0.6444 -0.0191 0.03781 fixed
count_birth_order3/4 0.01186 0.01571 0.7549 -0.01893 0.04265 fixed
count_birth_order4/4 0.006457 0.01706 0.3785 -0.02698 0.0399 fixed
count_birth_order1/5 0.0004113 0.01665 0.0247 -0.03223 0.03305 fixed
count_birth_order2/5 -0.01248 0.01707 -0.731 -0.04594 0.02098 fixed
count_birth_order3/5 -0.004494 0.01756 -0.2559 -0.03891 0.02992 fixed
count_birth_order4/5 -0.01346 0.01788 -0.753 -0.0485 0.02157 fixed
count_birth_order5/5 -0.002927 0.01815 -0.1613 -0.0385 0.03264 fixed
count_birth_order1/5+ 0.009889 0.01503 0.6581 -0.01956 0.03934 fixed
count_birth_order2/5+ 0.007599 0.01587 0.4788 -0.0235 0.0387 fixed
count_birth_order3/5+ 0.02187 0.01529 1.43 -0.008095 0.05183 fixed
count_birth_order4/5+ -0.009521 0.01485 -0.6411 -0.03863 0.01959 fixed
count_birth_order5/5+ 0.004109 0.01572 0.2613 -0.02671 0.03493 fixed
count_birth_order5+/5+ 0.00764 0.01189 0.6425 -0.01566 0.03094 fixed
sd_(Intercept).mother_pidlink 0.05089 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1503 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -3126 -3058 1574 -3148 NA NA NA
12 -3125 -3050 1574 -3149 0.3851 1 0.5349
16 -3121 -3021 1576 -3153 3.745 4 0.4416
26 -3108 -2946 1580 -3160 6.999 10 0.7256

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.01066 0.1071 -0.09954 -0.2206 0.1993 fixed
poly(age, 3, raw = TRUE)1 0.0003278 0.01151 0.02848 -0.02223 0.02289 fixed
poly(age, 3, raw = TRUE)2 0.00003105 0.0003947 0.07867 -0.0007425 0.0008046 fixed
poly(age, 3, raw = TRUE)3 -0.0000008807 0.000004332 -0.2033 -0.000009371 0.000007609 fixed
male 0.02676 0.005405 4.952 0.01617 0.03736 fixed
sibling_count3 0.01389 0.008671 1.602 -0.003107 0.03088 fixed
sibling_count4 0.01566 0.009086 1.723 -0.002149 0.03347 fixed
sibling_count5 0.003031 0.01042 0.2908 -0.0174 0.02346 fixed
sibling_count5+ 0.01426 0.008852 1.611 -0.003091 0.03161 fixed
sd_(Intercept).mother_pidlink 0.05302 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1516 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.01296 0.1072 -0.1209 -0.223 0.1971 fixed
birth_order -0.001449 0.001711 -0.8467 -0.004802 0.001905 fixed
poly(age, 3, raw = TRUE)1 0.00075 0.01152 0.0651 -0.02183 0.02333 fixed
poly(age, 3, raw = TRUE)2 0.00001983 0.0003949 0.05021 -0.0007542 0.0007938 fixed
poly(age, 3, raw = TRUE)3 -0.0000008209 0.000004332 -0.1895 -0.000009312 0.00000767 fixed
male 0.02679 0.005405 4.957 0.0162 0.03738 fixed
sibling_count3 0.01463 0.008718 1.679 -0.002453 0.03172 fixed
sibling_count4 0.01724 0.009279 1.858 -0.0009443 0.03543 fixed
sibling_count5 0.005559 0.01085 0.5122 -0.01571 0.02683 fixed
sibling_count5+ 0.01948 0.01078 1.806 -0.00166 0.04061 fixed
sd_(Intercept).mother_pidlink 0.05347 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1514 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.01281 0.1072 -0.1194 -0.223 0.1974 fixed
poly(age, 3, raw = TRUE)1 0.0007721 0.01152 0.06702 -0.02181 0.02335 fixed
poly(age, 3, raw = TRUE)2 0.0000206 0.0003949 0.05216 -0.0007535 0.0007947 fixed
poly(age, 3, raw = TRUE)3 -0.000000849 0.000004334 -0.1959 -0.000009343 0.000007645 fixed
male 0.02683 0.005405 4.963 0.01623 0.03742 fixed
sibling_count3 0.01413 0.008893 1.588 -0.003304 0.03156 fixed
sibling_count4 0.01947 0.009615 2.025 0.0006293 0.03832 fixed
sibling_count5 0.009835 0.01131 0.8699 -0.01232 0.03199 fixed
sibling_count5+ 0.02072 0.01108 1.871 -0.0009908 0.04243 fixed
birth_order_nonlinear2 -0.007982 0.007046 -1.133 -0.02179 0.005828 fixed
birth_order_nonlinear3 -0.001416 0.008319 -0.1702 -0.01772 0.01489 fixed
birth_order_nonlinear4 -0.02045 0.01045 -1.957 -0.04093 0.00002855 fixed
birth_order_nonlinear5 -0.01796 0.01288 -1.395 -0.04321 0.007276 fixed
birth_order_nonlinear5+ -0.006144 0.0127 -0.4838 -0.03104 0.01875 fixed
sd_(Intercept).mother_pidlink 0.0534 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1514 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.01112 0.1077 -0.1032 -0.2222 0.2 fixed
poly(age, 3, raw = TRUE)1 0.0006036 0.01157 0.05216 -0.02208 0.02328 fixed
poly(age, 3, raw = TRUE)2 0.00002941 0.0003968 0.07411 -0.0007484 0.0008072 fixed
poly(age, 3, raw = TRUE)3 -0.0000009731 0.000004356 -0.2234 -0.00000951 0.000007564 fixed
male 0.02685 0.005417 4.956 0.01623 0.03747 fixed
count_birth_order2/2 -0.01103 0.01345 -0.8202 -0.03738 0.01533 fixed
count_birth_order1/3 0.007516 0.01142 0.658 -0.01487 0.0299 fixed
count_birth_order2/3 0.01549 0.0127 1.22 -0.009394 0.04038 fixed
count_birth_order3/3 0.009049 0.01367 0.6619 -0.01774 0.03584 fixed
count_birth_order1/4 0.02297 0.0134 1.714 -0.003294 0.04924 fixed
count_birth_order2/4 0.0009928 0.01407 0.07055 -0.02659 0.02857 fixed
count_birth_order3/4 0.01526 0.0146 1.045 -0.01335 0.04388 fixed
count_birth_order4/4 0.005978 0.01586 0.3769 -0.02511 0.03706 fixed
count_birth_order1/5 0.007985 0.01752 0.4559 -0.02634 0.04231 fixed
count_birth_order2/5 -0.005466 0.01954 -0.2797 -0.04377 0.03284 fixed
count_birth_order3/5 0.01416 0.01906 0.7431 -0.02319 0.05151 fixed
count_birth_order4/5 -0.01167 0.01834 -0.636 -0.04762 0.02429 fixed
count_birth_order5/5 -0.009098 0.01948 -0.467 -0.04728 0.02909 fixed
count_birth_order1/5+ 0.02622 0.01707 1.536 -0.007235 0.05967 fixed
count_birth_order2/5+ 0.01201 0.01725 0.6963 -0.0218 0.04582 fixed
count_birth_order3/5+ 0.02036 0.01647 1.236 -0.01192 0.05263 fixed
count_birth_order4/5+ -0.009475 0.01653 -0.573 -0.04188 0.02293 fixed
count_birth_order5/5+ 0.00164 0.01552 0.1056 -0.02879 0.03207 fixed
count_birth_order5+/5+ 0.01344 0.01213 1.108 -0.01033 0.0372 fixed
sd_(Intercept).mother_pidlink 0.05348 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1516 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -2965 -2897 1493 -2987 NA NA NA
12 -2963 -2889 1494 -2987 0.7092 1 0.3997
16 -2960 -2861 1496 -2992 4.896 4 0.2981
26 -2944 -2783 1498 -2996 3.888 10 0.9523

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Sector_Social services

birthorder <- birthorder %>% mutate(outcome = `Sector_Social services`)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.02077 0.05104 0.407 -0.07926 0.1208 fixed
poly(age, 3, raw = TRUE)1 0.003093 0.004566 0.6775 -0.005855 0.01204 fixed
poly(age, 3, raw = TRUE)2 -0.00006539 0.0001278 -0.5117 -0.0003159 0.0001851 fixed
poly(age, 3, raw = TRUE)3 0.0000003566 0.000001132 0.3151 -0.000001862 0.000002575 fixed
male 0.001089 0.004758 0.2289 -0.008236 0.01041 fixed
sibling_count3 -0.0007703 0.009727 -0.07919 -0.01983 0.01829 fixed
sibling_count4 -0.006216 0.009798 -0.6344 -0.02542 0.01299 fixed
sibling_count5 -0.01513 0.01016 -1.489 -0.03504 0.004785 fixed
sibling_count5+ -0.005755 0.007926 -0.7261 -0.02129 0.00978 fixed
sd_(Intercept).mother_pidlink 0.05552 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2252 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.02068 0.05103 0.4051 -0.07935 0.1207 fixed
birth_order -0.0014 0.0009492 -1.475 -0.003261 0.0004599 fixed
poly(age, 3, raw = TRUE)1 0.003478 0.004573 0.7606 -0.005484 0.01244 fixed
poly(age, 3, raw = TRUE)2 -0.00007899 0.0001281 -0.6165 -0.0003301 0.0001721 fixed
poly(age, 3, raw = TRUE)3 0.0000004777 0.000001135 0.421 -0.000001746 0.000002702 fixed
male 0.001123 0.004758 0.2361 -0.008202 0.01045 fixed
sibling_count3 -0.0003848 0.00973 -0.03955 -0.01946 0.01869 fixed
sibling_count4 -0.005314 0.009817 -0.5413 -0.02456 0.01393 fixed
sibling_count5 -0.01359 0.01021 -1.331 -0.03361 0.006424 fixed
sibling_count5+ -0.0007645 0.008617 -0.08872 -0.01765 0.01612 fixed
sd_(Intercept).mother_pidlink 0.0556 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2251 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.01725 0.05113 0.3374 -0.08296 0.1175 fixed
poly(age, 3, raw = TRUE)1 0.003601 0.004578 0.7867 -0.005371 0.01257 fixed
poly(age, 3, raw = TRUE)2 -0.00008152 0.0001282 -0.6357 -0.0003328 0.0001698 fixed
poly(age, 3, raw = TRUE)3 0.0000004881 0.000001135 0.4299 -0.000001737 0.000002713 fixed
male 0.001126 0.004759 0.2366 -0.008201 0.01045 fixed
sibling_count3 0.001311 0.009885 0.1326 -0.01806 0.02069 fixed
sibling_count4 -0.003161 0.01011 -0.3127 -0.02297 0.01665 fixed
sibling_count5 -0.01113 0.01059 -1.05 -0.03188 0.009633 fixed
sibling_count5+ 0.001919 0.009053 0.212 -0.01582 0.01966 fixed
birth_order_nonlinear2 0.00005417 0.006983 0.007757 -0.01363 0.01374 fixed
birth_order_nonlinear3 -0.009658 0.008114 -1.19 -0.02556 0.006246 fixed
birth_order_nonlinear4 -0.007049 0.00912 -0.7729 -0.02492 0.01083 fixed
birth_order_nonlinear5 -0.009026 0.01024 -0.8813 -0.0291 0.01105 fixed
birth_order_nonlinear5+ -0.01269 0.00842 -1.507 -0.02919 0.003817 fixed
sd_(Intercept).mother_pidlink 0.05552 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2252 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.01213 0.05132 0.2363 -0.08846 0.1127 fixed
poly(age, 3, raw = TRUE)1 0.00332 0.004583 0.7244 -0.005662 0.0123 fixed
poly(age, 3, raw = TRUE)2 -0.00007672 0.0001283 -0.5978 -0.0003282 0.0001748 fixed
poly(age, 3, raw = TRUE)3 0.0000004767 0.000001136 0.4196 -0.00000175 0.000002703 fixed
male 0.001165 0.004761 0.2448 -0.008166 0.0105 fixed
count_birth_order2/2 0.02407 0.01392 1.729 -0.003215 0.05136 fixed
count_birth_order1/3 0.0107 0.01329 0.8051 -0.01535 0.03675 fixed
count_birth_order2/3 0.002195 0.01479 0.1485 -0.02678 0.03117 fixed
count_birth_order3/3 0.01111 0.01623 0.6843 -0.02071 0.04293 fixed
count_birth_order1/4 0.01602 0.01463 1.095 -0.01265 0.04469 fixed
count_birth_order2/4 0.003873 0.01569 0.2468 -0.02688 0.03463 fixed
count_birth_order3/4 -0.01074 0.01655 -0.6491 -0.04318 0.02169 fixed
count_birth_order4/4 -0.006686 0.01782 -0.3752 -0.04161 0.02824 fixed
count_birth_order1/5 -0.00212 0.0165 -0.1285 -0.03446 0.03022 fixed
count_birth_order2/5 -0.02531 0.01752 -1.445 -0.05964 0.00903 fixed
count_birth_order3/5 -0.00383 0.01843 -0.2078 -0.03996 0.0323 fixed
count_birth_order4/5 -0.003889 0.01946 -0.1998 -0.04204 0.03426 fixed
count_birth_order5/5 0.006752 0.0195 0.3463 -0.03146 0.04496 fixed
count_birth_order1/5+ 0.01512 0.01277 1.184 -0.009901 0.04014 fixed
count_birth_order2/5+ 0.01639 0.01327 1.235 -0.009619 0.04241 fixed
count_birth_order3/5+ -0.002849 0.01314 -0.2169 -0.0286 0.0229 fixed
count_birth_order4/5+ 0.004882 0.01287 0.3794 -0.02034 0.03011 fixed
count_birth_order5/5+ -0.002996 0.01301 -0.2303 -0.02849 0.0225 fixed
count_birth_order5+/5+ -0.001484 0.01046 -0.1419 -0.02198 0.01901 fixed
sd_(Intercept).mother_pidlink 0.05579 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2251 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -837.4 -758.3 429.7 -859.4 NA NA NA
12 -837.6 -751.3 430.8 -861.6 2.178 1 0.14
16 -830.8 -715.8 431.4 -862.8 1.23 4 0.8731
26 -820.8 -634 436.4 -872.8 10.06 10 0.4357

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.2203 0.156 -1.412 -0.526 0.08544 fixed
poly(age, 3, raw = TRUE)1 0.02853 0.01673 1.705 -0.004262 0.06132 fixed
poly(age, 3, raw = TRUE)2 -0.0009087 0.0005731 -1.586 -0.002032 0.0002145 fixed
poly(age, 3, raw = TRUE)3 0.000009383 0.000006282 1.494 -0.000002928 0.00002169 fixed
male 0.00001633 0.007889 0.00207 -0.01545 0.01548 fixed
sibling_count3 -0.005898 0.01284 -0.4594 -0.03106 0.01926 fixed
sibling_count4 -0.01841 0.01322 -1.393 -0.04432 0.007494 fixed
sibling_count5 -0.001773 0.01468 -0.1208 -0.03054 0.02699 fixed
sibling_count5+ -0.007312 0.01267 -0.5772 -0.03214 0.01752 fixed
sd_(Intercept).mother_pidlink 0.04285 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2318 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.2228 0.156 -1.428 -0.5285 0.08295 fixed
birth_order -0.002547 0.002424 -1.051 -0.007297 0.002203 fixed
poly(age, 3, raw = TRUE)1 0.02909 0.01674 1.738 -0.003718 0.0619 fixed
poly(age, 3, raw = TRUE)2 -0.000922 0.0005732 -1.608 -0.002046 0.0002015 fixed
poly(age, 3, raw = TRUE)3 0.000009422 0.000006282 1.5 -0.00000289 0.00002173 fixed
male 0.0001252 0.00789 0.01587 -0.01534 0.01559 fixed
sibling_count3 -0.004632 0.01289 -0.3593 -0.0299 0.02064 fixed
sibling_count4 -0.01563 0.01348 -1.16 -0.04205 0.01078 fixed
sibling_count5 0.002801 0.01531 0.183 -0.0272 0.0328 fixed
sibling_count5+ 0.001766 0.01533 0.1152 -0.02828 0.03181 fixed
sd_(Intercept).mother_pidlink 0.04259 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2319 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.2202 0.1562 -1.41 -0.5263 0.0859 fixed
poly(age, 3, raw = TRUE)1 0.02881 0.01675 1.72 -0.004019 0.06163 fixed
poly(age, 3, raw = TRUE)2 -0.0009102 0.0005736 -1.587 -0.002035 0.0002141 fixed
poly(age, 3, raw = TRUE)3 0.000009273 0.000006287 1.475 -0.00000305 0.0000216 fixed
male -0.00008511 0.007895 -0.01078 -0.01556 0.01539 fixed
sibling_count3 -0.003592 0.01316 -0.2729 -0.02939 0.0222 fixed
sibling_count4 -0.01523 0.01401 -1.088 -0.04268 0.01222 fixed
sibling_count5 0.001798 0.01608 0.1118 -0.02972 0.03331 fixed
sibling_count5+ 0.00193 0.01577 0.1224 -0.02898 0.03284 fixed
birth_order_nonlinear2 -0.0118 0.01054 -1.119 -0.03245 0.00886 fixed
birth_order_nonlinear3 -0.01021 0.01236 -0.8264 -0.03443 0.01401 fixed
birth_order_nonlinear4 -0.006922 0.01507 -0.4593 -0.03646 0.02262 fixed
birth_order_nonlinear5 -0.00769 0.01848 -0.4161 -0.04391 0.02853 fixed
birth_order_nonlinear5+ -0.02267 0.01806 -1.255 -0.05807 0.01273 fixed
sd_(Intercept).mother_pidlink 0.04289 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2319 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.2337 0.1567 -1.491 -0.5409 0.07343 fixed
poly(age, 3, raw = TRUE)1 0.03014 0.01681 1.793 -0.002808 0.06308 fixed
poly(age, 3, raw = TRUE)2 -0.0009591 0.0005759 -1.665 -0.002088 0.0001696 fixed
poly(age, 3, raw = TRUE)3 0.000009839 0.000006313 1.558 -0.000002535 0.00002221 fixed
male -0.00011 0.007904 -0.01392 -0.0156 0.01538 fixed
count_birth_order2/2 -0.005228 0.02055 -0.2543 -0.04551 0.03506 fixed
count_birth_order1/3 -0.004104 0.01708 -0.2403 -0.03758 0.02937 fixed
count_birth_order2/3 -0.03045 0.01899 -1.604 -0.06768 0.006769 fixed
count_birth_order3/3 0.01606 0.0207 0.7762 -0.0245 0.05663 fixed
count_birth_order1/4 -0.01085 0.0196 -0.5534 -0.04925 0.02756 fixed
count_birth_order2/4 -0.0261 0.02078 -1.256 -0.06683 0.01462 fixed
count_birth_order3/4 -0.0325 0.02155 -1.508 -0.07474 0.009748 fixed
count_birth_order4/4 -0.01051 0.02308 -0.4555 -0.05574 0.03472 fixed
count_birth_order1/5 0.0168 0.02582 0.6509 -0.0338 0.06741 fixed
count_birth_order2/5 -0.01284 0.02786 -0.4608 -0.06744 0.04177 fixed
count_birth_order3/5 -0.00643 0.02649 -0.2428 -0.05834 0.04548 fixed
count_birth_order4/5 0.007478 0.02584 0.2894 -0.04317 0.05812 fixed
count_birth_order5/5 -0.02538 0.0272 -0.9331 -0.07868 0.02793 fixed
count_birth_order1/5+ 0.003552 0.02444 0.1453 -0.04435 0.05146 fixed
count_birth_order2/5+ 0.02602 0.02465 1.056 -0.02229 0.07433 fixed
count_birth_order3/5+ -0.03433 0.02406 -1.427 -0.08148 0.01282 fixed
count_birth_order4/5+ -0.02051 0.02356 -0.8706 -0.06667 0.02566 fixed
count_birth_order5/5+ 0.009459 0.02251 0.4202 -0.03466 0.05358 fixed
count_birth_order5+/5+ -0.01853 0.01747 -1.061 -0.05276 0.0157 fixed
sd_(Intercept).mother_pidlink 0.04403 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2317 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -179.5 -111.2 100.8 -201.5 NA NA NA
12 -178.6 -104.1 101.3 -202.6 1.109 1 0.2924
16 -171.9 -72.52 101.9 -203.9 1.257 4 0.8687
26 -163.7 -2.307 107.9 -215.7 11.87 10 0.2936

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.2294 0.1558 -1.472 -0.5348 0.07603 fixed
poly(age, 3, raw = TRUE)1 0.02951 0.01673 1.764 -0.003278 0.0623 fixed
poly(age, 3, raw = TRUE)2 -0.0009341 0.0005732 -1.63 -0.002058 0.0001894 fixed
poly(age, 3, raw = TRUE)3 0.000009572 0.000006284 1.523 -0.000002745 0.00002189 fixed
male 0.0005569 0.00789 0.07059 -0.01491 0.01602 fixed
sibling_count3 -0.009068 0.01413 -0.6418 -0.03676 0.01863 fixed
sibling_count4 -0.01734 0.01423 -1.218 -0.04522 0.01055 fixed
sibling_count5 -0.01372 0.01495 -0.9178 -0.04302 0.01558 fixed
sibling_count5+ -0.006611 0.01311 -0.5042 -0.03231 0.01909 fixed
sd_(Intercept).mother_pidlink 0.04281 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2327 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.2319 0.1558 -1.488 -0.5373 0.07352 fixed
birth_order -0.003014 0.002142 -1.407 -0.007213 0.001185 fixed
poly(age, 3, raw = TRUE)1 0.03012 0.01673 1.8 -0.002678 0.06292 fixed
poly(age, 3, raw = TRUE)2 -0.0009479 0.0005732 -1.654 -0.002071 0.0001756 fixed
poly(age, 3, raw = TRUE)3 0.000009596 0.000006283 1.527 -0.00000272 0.00002191 fixed
male 0.0006874 0.00789 0.08713 -0.01478 0.01615 fixed
sibling_count3 -0.00756 0.01417 -0.5336 -0.03533 0.02021 fixed
sibling_count4 -0.01432 0.01439 -0.9953 -0.04251 0.01388 fixed
sibling_count5 -0.008675 0.01537 -0.5644 -0.0388 0.02145 fixed
sibling_count5+ 0.003984 0.01511 0.2636 -0.02564 0.03361 fixed
sd_(Intercept).mother_pidlink 0.04245 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2327 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.2256 0.156 -1.446 -0.5314 0.08022 fixed
poly(age, 3, raw = TRUE)1 0.02936 0.01675 1.753 -0.003465 0.06218 fixed
poly(age, 3, raw = TRUE)2 -0.0009206 0.0005737 -1.605 -0.002045 0.0002038 fixed
poly(age, 3, raw = TRUE)3 0.000009293 0.00000629 1.477 -0.000003036 0.00002162 fixed
male 0.0004118 0.007895 0.05216 -0.01506 0.01589 fixed
sibling_count3 -0.006194 0.01442 -0.4295 -0.03446 0.02207 fixed
sibling_count4 -0.01375 0.01483 -0.9267 -0.04282 0.01533 fixed
sibling_count5 -0.007203 0.01608 -0.4478 -0.03873 0.02432 fixed
sibling_count5+ 0.004556 0.01559 0.2923 -0.026 0.03511 fixed
birth_order_nonlinear2 -0.01058 0.0107 -0.9893 -0.03155 0.01038 fixed
birth_order_nonlinear3 -0.01241 0.01246 -0.9963 -0.03683 0.01201 fixed
birth_order_nonlinear4 -0.007174 0.01475 -0.4865 -0.03607 0.02173 fixed
birth_order_nonlinear5 -0.02097 0.01796 -1.168 -0.05617 0.01423 fixed
birth_order_nonlinear5+ -0.02246 0.01639 -1.371 -0.05457 0.009655 fixed
sd_(Intercept).mother_pidlink 0.04251 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2328 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.2352 0.1565 -1.503 -0.5418 0.07143 fixed
poly(age, 3, raw = TRUE)1 0.03016 0.0168 1.796 -0.002758 0.06308 fixed
poly(age, 3, raw = TRUE)2 -0.0009528 0.0005756 -1.655 -0.002081 0.0001753 fixed
poly(age, 3, raw = TRUE)3 0.000009682 0.000006311 1.534 -0.000002688 0.00002205 fixed
male 0.0004338 0.00791 0.05484 -0.01507 0.01594 fixed
count_birth_order2/2 -0.00005092 0.02262 -0.002251 -0.04438 0.04428 fixed
count_birth_order1/3 -0.008272 0.01896 -0.4363 -0.04543 0.02889 fixed
count_birth_order2/3 -0.0263 0.02066 -1.273 -0.06679 0.01418 fixed
count_birth_order3/3 0.01414 0.0229 0.6175 -0.03075 0.05903 fixed
count_birth_order1/4 -0.002246 0.02058 -0.1092 -0.04258 0.03808 fixed
count_birth_order2/4 -0.03167 0.02169 -1.46 -0.07418 0.01084 fixed
count_birth_order3/4 -0.02752 0.02348 -1.172 -0.07354 0.01851 fixed
count_birth_order4/4 -0.008276 0.02551 -0.3244 -0.05828 0.04172 fixed
count_birth_order1/5 0.000694 0.02488 0.0279 -0.04807 0.04945 fixed
count_birth_order2/5 -0.01059 0.02552 -0.4151 -0.0606 0.03942 fixed
count_birth_order3/5 -0.01189 0.02626 -0.4526 -0.06336 0.03958 fixed
count_birth_order4/5 -0.01932 0.02674 -0.7227 -0.07173 0.03308 fixed
count_birth_order5/5 -0.02963 0.02715 -1.091 -0.08285 0.02359 fixed
count_birth_order1/5+ 0.01111 0.02243 0.4953 -0.03286 0.05508 fixed
count_birth_order2/5+ 0.02061 0.02372 0.8689 -0.02588 0.0671 fixed
count_birth_order3/5+ -0.02985 0.02286 -1.306 -0.07465 0.01495 fixed
count_birth_order4/5+ 0.0001227 0.02221 0.005523 -0.04341 0.04365 fixed
count_birth_order5/5+ -0.009344 0.02353 -0.3971 -0.05546 0.03677 fixed
count_birth_order5+/5+ -0.01417 0.01765 -0.8029 -0.04877 0.02042 fixed
sd_(Intercept).mother_pidlink 0.04264 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2328 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -155.6 -87.25 88.81 -177.6 NA NA NA
12 -155.6 -81.02 89.8 -179.6 1.987 1 0.1587
16 -148.5 -49.03 90.24 -180.5 0.8671 4 0.9292
26 -136.8 24.85 94.38 -188.8 8.28 10 0.6015

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.1795 0.155 -1.158 -0.4832 0.1242 fixed
poly(age, 3, raw = TRUE)1 0.02407 0.01665 1.446 -0.008562 0.0567 fixed
poly(age, 3, raw = TRUE)2 -0.0007797 0.0005708 -1.366 -0.001898 0.000339 fixed
poly(age, 3, raw = TRUE)3 0.000008248 0.000006263 1.317 -0.000004027 0.00002052 fixed
male 0.0003739 0.007836 0.04772 -0.01498 0.01573 fixed
sibling_count3 -0.003764 0.01234 -0.3051 -0.02794 0.02041 fixed
sibling_count4 -0.01422 0.01288 -1.105 -0.03946 0.01101 fixed
sibling_count5 0.003677 0.01466 0.2508 -0.02506 0.03241 fixed
sibling_count5+ 0.00007903 0.01245 0.006349 -0.02432 0.02448 fixed
sd_(Intercept).mother_pidlink 0.03288 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2297 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.182 0.155 -1.174 -0.4858 0.1218 fixed
birth_order -0.001948 0.002453 -0.7941 -0.006756 0.00286 fixed
poly(age, 3, raw = TRUE)1 0.02457 0.01666 1.474 -0.00809 0.05722 fixed
poly(age, 3, raw = TRUE)2 -0.0007923 0.000571 -1.387 -0.001911 0.000327 fixed
poly(age, 3, raw = TRUE)3 0.000008303 0.000006264 1.326 -0.000003973 0.00002058 fixed
male 0.0004034 0.007837 0.05148 -0.01496 0.01576 fixed
sibling_count3 -0.002778 0.0124 -0.2241 -0.02708 0.02152 fixed
sibling_count4 -0.01213 0.01314 -0.9225 -0.03789 0.01364 fixed
sibling_count5 0.007049 0.01526 0.4619 -0.02286 0.03696 fixed
sibling_count5+ 0.007013 0.0152 0.4613 -0.02278 0.03681 fixed
sd_(Intercept).mother_pidlink 0.03272 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2297 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.18 0.1552 -1.16 -0.4841 0.1242 fixed
poly(age, 3, raw = TRUE)1 0.02441 0.01667 1.464 -0.008266 0.05708 fixed
poly(age, 3, raw = TRUE)2 -0.0007846 0.0005714 -1.373 -0.001905 0.0003353 fixed
poly(age, 3, raw = TRUE)3 0.000008199 0.000006269 1.308 -0.000004088 0.00002049 fixed
male 0.000236 0.007841 0.0301 -0.01513 0.0156 fixed
sibling_count3 -0.002512 0.01268 -0.1982 -0.02736 0.02233 fixed
sibling_count4 -0.01313 0.01367 -0.9604 -0.03992 0.01366 fixed
sibling_count5 0.004467 0.01598 0.2795 -0.02686 0.03579 fixed
sibling_count5+ 0.006556 0.01567 0.4184 -0.02416 0.03727 fixed
birth_order_nonlinear2 -0.01101 0.01035 -1.064 -0.03129 0.009268 fixed
birth_order_nonlinear3 -0.006014 0.0122 -0.4931 -0.02992 0.01789 fixed
birth_order_nonlinear4 -0.001251 0.0153 -0.08179 -0.03123 0.02873 fixed
birth_order_nonlinear5 -0.004343 0.01886 -0.2303 -0.0413 0.03262 fixed
birth_order_nonlinear5+ -0.01995 0.01833 -1.088 -0.05587 0.01597 fixed
sd_(Intercept).mother_pidlink 0.03303 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2298 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.1869 0.1557 -1.201 -0.492 0.1182 fixed
poly(age, 3, raw = TRUE)1 0.02514 0.01673 1.503 -0.00764 0.05793 fixed
poly(age, 3, raw = TRUE)2 -0.000813 0.0005736 -1.418 -0.001937 0.0003111 fixed
poly(age, 3, raw = TRUE)3 0.000008544 0.000006294 1.357 -0.000003793 0.00002088 fixed
male 0.0004866 0.007851 0.06198 -0.0149 0.01588 fixed
count_birth_order2/2 -0.008769 0.01971 -0.4448 -0.04741 0.02987 fixed
count_birth_order1/3 -0.004317 0.01651 -0.2614 -0.03669 0.02805 fixed
count_birth_order2/3 -0.02721 0.01838 -1.481 -0.06322 0.008806 fixed
count_birth_order3/3 0.01517 0.0198 0.7662 -0.02364 0.05398 fixed
count_birth_order1/4 -0.01353 0.01938 -0.6981 -0.05153 0.02446 fixed
count_birth_order2/4 -0.02109 0.02038 -1.035 -0.06103 0.01884 fixed
count_birth_order3/4 -0.02761 0.02115 -1.306 -0.06907 0.01384 fixed
count_birth_order4/4 -0.003205 0.02298 -0.1395 -0.04824 0.04183 fixed
count_birth_order1/5 0.01094 0.02535 0.4314 -0.03874 0.06062 fixed
count_birth_order2/5 -0.01583 0.02833 -0.559 -0.07135 0.03968 fixed
count_birth_order3/5 0.007652 0.02763 0.277 -0.04649 0.0618 fixed
count_birth_order4/5 0.01138 0.02659 0.4279 -0.04074 0.0635 fixed
count_birth_order5/5 -0.01412 0.02825 -0.4998 -0.0695 0.04126 fixed
count_birth_order1/5+ 0.01476 0.02469 0.5976 -0.03364 0.06315 fixed
count_birth_order2/5+ 0.02997 0.02499 1.199 -0.019 0.07894 fixed
count_birth_order3/5+ -0.02888 0.02386 -1.211 -0.07565 0.01788 fixed
count_birth_order4/5+ -0.01081 0.02398 -0.451 -0.0578 0.03618 fixed
count_birth_order5/5+ 0.01124 0.02251 0.4996 -0.03287 0.05536 fixed
count_birth_order5+/5+ -0.01263 0.01737 -0.7269 -0.04667 0.02142 fixed
sd_(Intercept).mother_pidlink 0.03417 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2296 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -290.8 -222.7 156.4 -312.8 NA NA NA
12 -289.4 -215.1 156.7 -313.4 0.6335 1 0.4261
16 -283 -183.9 157.5 -315 1.525 4 0.8222
26 -273 -112.1 162.5 -325 10.08 10 0.4334

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Sector_Transportation, storage and communications

birthorder <- birthorder %>% mutate(outcome = `Sector_Transportation, storage and communications`)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.0652 0.04159 1.568 -0.01631 0.1467 fixed
poly(age, 3, raw = TRUE)1 -0.003102 0.003721 -0.8336 -0.01039 0.004191 fixed
poly(age, 3, raw = TRUE)2 0.00007889 0.0001042 0.7573 -0.0001253 0.0002831 fixed
poly(age, 3, raw = TRUE)3 -0.0000005583 0.0000009226 -0.6052 -0.000002367 0.00000125 fixed
male -0.01246 0.003874 -3.218 -0.02006 -0.004872 fixed
sibling_count3 0.008001 0.007939 1.008 -0.00756 0.02356 fixed
sibling_count4 0.009834 0.008 1.229 -0.005846 0.02551 fixed
sibling_count5 0.02201 0.008298 2.653 0.005748 0.03828 fixed
sibling_count5+ 0.01792 0.00647 2.769 0.005236 0.0306 fixed
sd_(Intercept).mother_pidlink 0.04837 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1826 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.06522 0.04159 1.568 -0.01629 0.1467 fixed
birth_order 0.0004447 0.0007746 0.5741 -0.001073 0.001963 fixed
poly(age, 3, raw = TRUE)1 -0.003224 0.003727 -0.865 -0.01053 0.004081 fixed
poly(age, 3, raw = TRUE)2 0.00008323 0.0001044 0.7969 -0.0001215 0.0002879 fixed
poly(age, 3, raw = TRUE)3 -0.000000597 0.0000009251 -0.6454 -0.00000241 0.000001216 fixed
male -0.01247 0.003874 -3.22 -0.02007 -0.004883 fixed
sibling_count3 0.007879 0.007942 0.992 -0.007687 0.02345 fixed
sibling_count4 0.00955 0.008016 1.191 -0.006161 0.02526 fixed
sibling_count5 0.02153 0.008341 2.581 0.005178 0.03788 fixed
sibling_count5+ 0.01634 0.007032 2.323 0.002554 0.03012 fixed
sd_(Intercept).mother_pidlink 0.04838 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1826 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.06535 0.04165 1.569 -0.01629 0.147 fixed
poly(age, 3, raw = TRUE)1 -0.003216 0.00373 -0.8623 -0.01053 0.004095 fixed
poly(age, 3, raw = TRUE)2 0.0000849 0.0001045 0.8124 -0.0001199 0.0002897 fixed
poly(age, 3, raw = TRUE)3 -0.0000006214 0.0000009255 -0.6715 -0.000002435 0.000001192 fixed
male -0.01255 0.003874 -3.241 -0.02015 -0.004962 fixed
sibling_count3 0.007229 0.008066 0.8962 -0.008581 0.02304 fixed
sibling_count4 0.01056 0.008249 1.28 -0.005609 0.02673 fixed
sibling_count5 0.02083 0.008646 2.409 0.00388 0.03777 fixed
sibling_count5+ 0.01412 0.007382 1.913 -0.0003453 0.02859 fixed
birth_order_nonlinear2 -0.001514 0.005678 -0.2667 -0.01264 0.009615 fixed
birth_order_nonlinear3 0.003304 0.006598 0.5007 -0.009628 0.01623 fixed
birth_order_nonlinear4 -0.007857 0.007416 -1.06 -0.02239 0.006677 fixed
birth_order_nonlinear5 0.01102 0.008329 1.324 -0.005301 0.02735 fixed
birth_order_nonlinear5+ 0.007088 0.006857 1.034 -0.006351 0.02053 fixed
sd_(Intercept).mother_pidlink 0.04841 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1826 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.06713 0.04182 1.605 -0.01484 0.1491 fixed
poly(age, 3, raw = TRUE)1 -0.003226 0.003735 -0.8637 -0.01055 0.004095 fixed
poly(age, 3, raw = TRUE)2 0.00008514 0.0001046 0.8139 -0.0001199 0.0002902 fixed
poly(age, 3, raw = TRUE)3 -0.0000006217 0.0000009261 -0.6713 -0.000002437 0.000001193 fixed
male -0.01247 0.003877 -3.216 -0.02007 -0.004869 fixed
count_birth_order2/2 -0.006295 0.01132 -0.556 -0.02849 0.0159 fixed
count_birth_order1/3 0.003135 0.01083 0.2896 -0.01808 0.02435 fixed
count_birth_order2/3 0.00541 0.01204 0.4493 -0.01819 0.02901 fixed
count_birth_order3/3 0.01109 0.01322 0.8389 -0.01482 0.03701 fixed
count_birth_order1/4 0.01225 0.01192 1.028 -0.0111 0.03561 fixed
count_birth_order2/4 0.00659 0.01278 0.5156 -0.01846 0.03164 fixed
count_birth_order3/4 0.01458 0.01348 1.082 -0.01184 0.041 fixed
count_birth_order4/4 -0.007248 0.01451 -0.4995 -0.03569 0.02119 fixed
count_birth_order1/5 0.004284 0.01344 0.3188 -0.02206 0.03063 fixed
count_birth_order2/5 0.02675 0.01427 1.875 -0.001216 0.05471 fixed
count_birth_order3/5 0.0287 0.01501 1.912 -0.0007223 0.05813 fixed
count_birth_order4/5 0.008712 0.01585 0.5497 -0.02235 0.03978 fixed
count_birth_order5/5 0.03591 0.01588 2.261 0.004787 0.06703 fixed
count_birth_order1/5+ 0.0164 0.0104 1.577 -0.003981 0.03677 fixed
count_birth_order2/5+ 0.008682 0.01081 0.8033 -0.0125 0.02987 fixed
count_birth_order3/5+ 0.01101 0.0107 1.029 -0.00996 0.03198 fixed
count_birth_order4/5+ 0.008169 0.01048 0.7795 -0.01237 0.02871 fixed
count_birth_order5/5+ 0.02166 0.01059 2.044 0.0008929 0.04242 fixed
count_birth_order5+/5+ 0.01945 0.008525 2.281 0.00274 0.03616 fixed
sd_(Intercept).mother_pidlink 0.04833 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1826 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -4846 -4767 2434 -4868 NA NA NA
12 -4844 -4758 2434 -4868 0.33 1 0.5656
16 -4842 -4727 2437 -4874 5.859 4 0.2099
26 -4827 -4640 2439 -4879 4.714 10 0.9094

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.147 0.1226 1.199 -0.09325 0.3872 fixed
poly(age, 3, raw = TRUE)1 -0.01108 0.01315 -0.8429 -0.03685 0.01468 fixed
poly(age, 3, raw = TRUE)2 0.0003627 0.0004503 0.8054 -0.0005199 0.001245 fixed
poly(age, 3, raw = TRUE)3 -0.000004015 0.000004936 -0.8135 -0.00001369 0.000005658 fixed
male -0.016 0.006199 -2.581 -0.02815 -0.003849 fixed
sibling_count3 0.004867 0.01008 0.4826 -0.0149 0.02463 fixed
sibling_count4 0.006221 0.01038 0.5992 -0.01413 0.02657 fixed
sibling_count5 0.0261 0.01153 2.264 0.003508 0.0487 fixed
sibling_count5+ 0.01793 0.009949 1.802 -0.001569 0.03743 fixed
sd_(Intercept).mother_pidlink 0.03293 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1823 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.1438 0.1225 1.173 -0.09643 0.3839 fixed
birth_order -0.00308 0.001903 -1.618 -0.006811 0.0006507 fixed
poly(age, 3, raw = TRUE)1 -0.01039 0.01315 -0.7901 -0.03616 0.01538 fixed
poly(age, 3, raw = TRUE)2 0.0003461 0.0004503 0.7686 -0.0005365 0.001229 fixed
poly(age, 3, raw = TRUE)3 -0.000003965 0.000004934 -0.8035 -0.00001364 0.000005707 fixed
male -0.01587 0.006198 -2.56 -0.02802 -0.003722 fixed
sibling_count3 0.006395 0.01013 0.6316 -0.01345 0.02624 fixed
sibling_count4 0.009575 0.01058 0.9046 -0.01117 0.03032 fixed
sibling_count5 0.03163 0.01202 2.632 0.008072 0.05518 fixed
sibling_count5+ 0.0289 0.01203 2.401 0.005311 0.05249 fixed
sd_(Intercept).mother_pidlink 0.03253 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1823 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.1472 0.1227 1.2 -0.09317 0.3876 fixed
poly(age, 3, raw = TRUE)1 -0.0108 0.01315 -0.821 -0.03658 0.01498 fixed
poly(age, 3, raw = TRUE)2 0.0003645 0.0004505 0.809 -0.0005185 0.001247 fixed
poly(age, 3, raw = TRUE)3 -0.00000422 0.000004938 -0.8547 -0.0000139 0.000005457 fixed
male -0.01611 0.006201 -2.598 -0.02826 -0.003957 fixed
sibling_count3 0.007144 0.01033 0.6914 -0.01311 0.0274 fixed
sibling_count4 0.01067 0.01099 0.971 -0.01087 0.03222 fixed
sibling_count5 0.03423 0.01262 2.713 0.009498 0.05897 fixed
sibling_count5+ 0.03208 0.01238 2.592 0.007826 0.05634 fixed
birth_order_nonlinear2 -0.01406 0.008281 -1.698 -0.03029 0.002167 fixed
birth_order_nonlinear3 -0.01005 0.009708 -1.035 -0.02908 0.008976 fixed
birth_order_nonlinear4 -0.01388 0.01184 -1.172 -0.03708 0.009328 fixed
birth_order_nonlinear5 -0.02517 0.01452 -1.734 -0.05363 0.003279 fixed
birth_order_nonlinear5+ -0.02715 0.01418 -1.915 -0.05495 0.0006431 fixed
sd_(Intercept).mother_pidlink 0.03234 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1824 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.1428 0.1231 1.16 -0.0985 0.3841 fixed
poly(age, 3, raw = TRUE)1 -0.01111 0.01321 -0.8416 -0.037 0.01477 fixed
poly(age, 3, raw = TRUE)2 0.0003788 0.0004524 0.8374 -0.0005079 0.001266 fixed
poly(age, 3, raw = TRUE)3 -0.000004428 0.00000496 -0.8929 -0.00001415 0.000005292 fixed
male -0.0164 0.00621 -2.64 -0.02857 -0.004225 fixed
count_birth_order2/2 0.006857 0.01616 0.4242 -0.02482 0.03853 fixed
count_birth_order1/3 0.009847 0.01342 0.7337 -0.01646 0.03615 fixed
count_birth_order2/3 -0.0002712 0.01492 -0.01818 -0.02952 0.02898 fixed
count_birth_order3/3 0.01219 0.01626 0.7498 -0.01968 0.04407 fixed
count_birth_order1/4 0.02105 0.0154 1.367 -0.009128 0.05123 fixed
count_birth_order2/4 0.009409 0.01633 0.5762 -0.02259 0.04141 fixed
count_birth_order3/4 0.0003186 0.01694 0.01881 -0.03288 0.03351 fixed
count_birth_order4/4 -0.001352 0.01813 -0.07459 -0.03689 0.03418 fixed
count_birth_order1/5 0.04666 0.02029 2.3 0.006901 0.08642 fixed
count_birth_order2/5 0.0226 0.02189 1.032 -0.02031 0.06551 fixed
count_birth_order3/5 0.03762 0.02081 1.808 -0.003166 0.07842 fixed
count_birth_order4/5 0.02517 0.0203 1.24 -0.01462 0.06497 fixed
count_birth_order5/5 0.009177 0.02137 0.4294 -0.03271 0.05106 fixed
count_birth_order1/5+ 0.06168 0.0192 3.212 0.02405 0.09932 fixed
count_birth_order2/5+ -0.002437 0.01937 -0.1258 -0.04039 0.03552 fixed
count_birth_order3/5+ 0.02147 0.0189 1.136 -0.01558 0.05852 fixed
count_birth_order4/5+ 0.03178 0.01851 1.717 -0.004493 0.06806 fixed
count_birth_order5/5+ 0.01785 0.01769 1.009 -0.01682 0.05252 fixed
count_birth_order5+/5+ 0.01175 0.01371 0.8568 -0.01513 0.03863 fixed
sd_(Intercept).mother_pidlink 0.0318 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1825 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -1951 -1883 986.6 -1973 NA NA NA
12 -1952 -1877 987.9 -1976 2.627 1 0.105
16 -1947 -1848 989.5 -1979 3.289 4 0.5106
26 -1936 -1774 993.8 -1988 8.507 10 0.5795

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.1348 0.122 1.105 -0.1043 0.3739 fixed
poly(age, 3, raw = TRUE)1 -0.01021 0.0131 -0.7795 -0.03588 0.01546 fixed
poly(age, 3, raw = TRUE)2 0.0003311 0.0004488 0.7378 -0.0005485 0.001211 fixed
poly(age, 3, raw = TRUE)3 -0.000003669 0.00000492 -0.7457 -0.00001331 0.000005974 fixed
male -0.0162 0.006178 -2.622 -0.0283 -0.004088 fixed
sibling_count3 0.009527 0.01106 0.8615 -0.01215 0.0312 fixed
sibling_count4 0.006535 0.01113 0.5869 -0.01529 0.02836 fixed
sibling_count5 0.0336 0.0117 2.872 0.01067 0.05653 fixed
sibling_count5+ 0.02293 0.01026 2.235 0.002817 0.04304 fixed
sd_(Intercept).mother_pidlink 0.03225 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1824 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.133 0.122 1.09 -0.1061 0.3721 fixed
birth_order -0.002138 0.001677 -1.275 -0.005425 0.001148 fixed
poly(age, 3, raw = TRUE)1 -0.009771 0.0131 -0.7458 -0.03545 0.01591 fixed
poly(age, 3, raw = TRUE)2 0.000321 0.0004488 0.7153 -0.0005586 0.001201 fixed
poly(age, 3, raw = TRUE)3 -0.00000365 0.00000492 -0.7418 -0.00001329 0.000005993 fixed
male -0.01611 0.006178 -2.607 -0.02822 -0.003998 fixed
sibling_count3 0.0106 0.01109 0.9557 -0.01114 0.03233 fixed
sibling_count4 0.008674 0.01126 0.7704 -0.01339 0.03074 fixed
sibling_count5 0.03718 0.01203 3.091 0.0136 0.06076 fixed
sibling_count5+ 0.03044 0.01183 2.573 0.007253 0.05362 fixed
sd_(Intercept).mother_pidlink 0.03218 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1824 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.1407 0.1221 1.152 -0.09868 0.38 fixed
poly(age, 3, raw = TRUE)1 -0.01045 0.01311 -0.797 -0.03613 0.01524 fixed
poly(age, 3, raw = TRUE)2 0.0003477 0.000449 0.7743 -0.0005324 0.001228 fixed
poly(age, 3, raw = TRUE)3 -0.00000398 0.000004923 -0.8085 -0.00001363 0.000005669 fixed
male -0.01637 0.006179 -2.649 -0.02848 -0.004261 fixed
sibling_count3 0.01284 0.01128 1.138 -0.00927 0.03496 fixed
sibling_count4 0.01034 0.0116 0.8908 -0.01241 0.03308 fixed
sibling_count5 0.03911 0.01258 3.108 0.01444 0.06377 fixed
sibling_count5+ 0.03291 0.01219 2.699 0.00901 0.05681 fixed
birth_order_nonlinear2 -0.015 0.008375 -1.791 -0.03141 0.001419 fixed
birth_order_nonlinear3 -0.01467 0.009754 -1.504 -0.03379 0.004447 fixed
birth_order_nonlinear4 -0.007381 0.01154 -0.6394 -0.03001 0.01524 fixed
birth_order_nonlinear5 -0.01601 0.01406 -1.138 -0.04356 0.01155 fixed
birth_order_nonlinear5+ -0.02288 0.01282 -1.785 -0.04802 0.002247 fixed
sd_(Intercept).mother_pidlink 0.03209 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1824 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.1241 0.1221 1.016 -0.1153 0.3635 fixed
poly(age, 3, raw = TRUE)1 -0.009842 0.01311 -0.7506 -0.03554 0.01586 fixed
poly(age, 3, raw = TRUE)2 0.000332 0.0004494 0.7389 -0.0005487 0.001213 fixed
poly(age, 3, raw = TRUE)3 -0.000003868 0.000004927 -0.785 -0.00001353 0.000005789 fixed
male -0.01598 0.006177 -2.587 -0.02809 -0.003875 fixed
count_birth_order2/2 0.0135 0.01767 0.7641 -0.02113 0.04813 fixed
count_birth_order1/3 0.01416 0.0148 0.9565 -0.01486 0.04317 fixed
count_birth_order2/3 0.00501 0.01613 0.3106 -0.0266 0.03662 fixed
count_birth_order3/3 0.0268 0.01788 1.499 -0.00825 0.06186 fixed
count_birth_order1/4 0.04209 0.01607 2.62 0.0106 0.07358 fixed
count_birth_order2/4 0.001262 0.01694 0.07454 -0.03193 0.03445 fixed
count_birth_order3/4 -0.01579 0.01834 -0.861 -0.05173 0.02015 fixed
count_birth_order4/4 0.001377 0.01992 0.06911 -0.03767 0.04042 fixed
count_birth_order1/5 0.01152 0.01943 0.5932 -0.02655 0.0496 fixed
count_birth_order2/5 0.04278 0.01992 2.147 0.003731 0.08183 fixed
count_birth_order3/5 0.05774 0.02051 2.816 0.01755 0.09793 fixed
count_birth_order4/5 0.05098 0.02088 2.441 0.01005 0.0919 fixed
count_birth_order5/5 0.03361 0.0212 1.585 -0.007949 0.07517 fixed
count_birth_order1/5+ 0.0718 0.01752 4.099 0.03746 0.1061 fixed
count_birth_order2/5+ 0.00797 0.01852 0.4303 -0.02833 0.04427 fixed
count_birth_order3/5+ 0.01289 0.01785 0.7224 -0.02209 0.04787 fixed
count_birth_order4/5+ 0.03653 0.01734 2.106 0.00254 0.07052 fixed
count_birth_order5/5+ 0.02583 0.01837 1.406 -0.01018 0.06183 fixed
count_birth_order5+/5+ 0.01946 0.01378 1.412 -0.007545 0.04646 fixed
sd_(Intercept).mother_pidlink 0.03092 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1822 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -1965 -1897 993.5 -1987 NA NA NA
12 -1965 -1890 994.3 -1989 1.631 1 0.2015
16 -1961 -1861 996.3 -1993 3.928 4 0.4158
26 -1966 -1805 1009 -2018 25.91 10 0.003867

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.1323 0.1239 1.068 -0.1105 0.375 fixed
poly(age, 3, raw = TRUE)1 -0.009582 0.01331 -0.72 -0.03566 0.0165 fixed
poly(age, 3, raw = TRUE)2 0.0003121 0.0004563 0.684 -0.0005822 0.001206 fixed
poly(age, 3, raw = TRUE)3 -0.000003541 0.000005007 -0.7072 -0.00001335 0.000006272 fixed
male -0.01515 0.006261 -2.42 -0.02742 -0.002879 fixed
sibling_count3 0.006555 0.009888 0.6629 -0.01283 0.02593 fixed
sibling_count4 0.0103 0.01033 0.9976 -0.00994 0.03055 fixed
sibling_count5 0.03153 0.01178 2.677 0.008448 0.05462 fixed
sibling_count5+ 0.02033 0.01 2.032 0.0007231 0.03993 fixed
sd_(Intercept).mother_pidlink 0.03537 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.182 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.1282 0.1239 1.035 -0.1145 0.371 fixed
birth_order -0.002884 0.001964 -1.469 -0.006734 0.0009649 fixed
poly(age, 3, raw = TRUE)1 -0.008817 0.01331 -0.6622 -0.03491 0.01728 fixed
poly(age, 3, raw = TRUE)2 0.0002926 0.0004564 0.6411 -0.0006019 0.001187 fixed
poly(age, 3, raw = TRUE)3 -0.00000345 0.000005006 -0.6891 -0.00001326 0.000006362 fixed
male -0.01511 0.00626 -2.414 -0.02738 -0.00284 fixed
sibling_count3 0.008016 0.009934 0.8069 -0.01145 0.02749 fixed
sibling_count4 0.01341 0.01054 1.272 -0.007251 0.03406 fixed
sibling_count5 0.03653 0.01225 2.98 0.01251 0.06054 fixed
sibling_count5+ 0.0306 0.0122 2.508 0.006687 0.05452 fixed
sd_(Intercept).mother_pidlink 0.03485 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1821 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.1321 0.124 1.066 -0.1109 0.3752 fixed
poly(age, 3, raw = TRUE)1 -0.009326 0.01332 -0.7001 -0.03543 0.01678 fixed
poly(age, 3, raw = TRUE)2 0.0003124 0.0004566 0.6841 -0.0005826 0.001207 fixed
poly(age, 3, raw = TRUE)3 -0.000003693 0.00000501 -0.7371 -0.00001351 0.000006127 fixed
male -0.01529 0.006263 -2.441 -0.02756 -0.00301 fixed
sibling_count3 0.008291 0.01015 0.8167 -0.01161 0.02819 fixed
sibling_count4 0.0134 0.01095 1.224 -0.008063 0.03487 fixed
sibling_count5 0.03807 0.01282 2.97 0.01295 0.06319 fixed
sibling_count5+ 0.03212 0.01257 2.556 0.007489 0.05675 fixed
birth_order_nonlinear2 -0.01056 0.008247 -1.281 -0.02672 0.005602 fixed
birth_order_nonlinear3 -0.007715 0.009724 -0.7933 -0.02677 0.01134 fixed
birth_order_nonlinear4 -0.009011 0.0122 -0.7386 -0.03292 0.0149 fixed
birth_order_nonlinear5 -0.02419 0.01504 -1.609 -0.05367 0.005283 fixed
birth_order_nonlinear5+ -0.0218 0.01465 -1.488 -0.05052 0.006923 fixed
sd_(Intercept).mother_pidlink 0.03468 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1822 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.121 0.1244 0.9726 -0.1228 0.3649 fixed
poly(age, 3, raw = TRUE)1 -0.008724 0.01337 -0.6525 -0.03493 0.01748 fixed
poly(age, 3, raw = TRUE)2 0.0002937 0.0004585 0.6407 -0.0006049 0.001192 fixed
poly(age, 3, raw = TRUE)3 -0.000003527 0.000005032 -0.7009 -0.00001339 0.000006335 fixed
male -0.0155 0.006273 -2.471 -0.0278 -0.003207 fixed
count_birth_order2/2 0.005899 0.01572 0.3752 -0.02492 0.03671 fixed
count_birth_order1/3 0.01002 0.0132 0.7592 -0.01585 0.03588 fixed
count_birth_order2/3 0.004641 0.01468 0.3161 -0.02413 0.03342 fixed
count_birth_order3/3 0.01141 0.01582 0.721 -0.0196 0.04241 fixed
count_birth_order1/4 0.02893 0.01549 1.868 -0.00143 0.05929 fixed
count_birth_order2/4 0.01144 0.01628 0.7029 -0.02046 0.04335 fixed
count_birth_order3/4 0.002006 0.0169 0.1187 -0.03111 0.03513 fixed
count_birth_order4/4 0.0008381 0.01836 0.04565 -0.03515 0.03682 fixed
count_birth_order1/5 0.04949 0.02025 2.443 0.00979 0.08918 fixed
count_birth_order2/5 0.02977 0.02263 1.316 -0.01458 0.07413 fixed
count_birth_order3/5 0.04936 0.02207 2.236 0.006101 0.09262 fixed
count_birth_order4/5 0.02171 0.02125 1.022 -0.01994 0.06335 fixed
count_birth_order5/5 0.01596 0.02257 0.7072 -0.02828 0.0602 fixed
count_birth_order1/5+ 0.04549 0.01974 2.305 0.006806 0.08417 fixed
count_birth_order2/5+ 0.001633 0.01997 0.08179 -0.0375 0.04076 fixed
count_birth_order3/5+ 0.02498 0.01907 1.31 -0.01239 0.06234 fixed
count_birth_order4/5+ 0.04811 0.01915 2.512 0.01057 0.08566 fixed
count_birth_order5/5+ 0.01544 0.01798 0.8588 -0.0198 0.05069 fixed
count_birth_order5+/5+ 0.01582 0.0139 1.138 -0.01143 0.04307 fixed
sd_(Intercept).mother_pidlink 0.03447 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.1823 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 -1911 -1843 966.6 -1933 NA NA NA
12 -1911 -1837 967.7 -1935 2.166 1 0.1411
16 -1905 -1806 968.7 -1937 1.891 4 0.7558
26 -1893 -1732 972.6 -1945 7.916 10 0.637

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Smoking Behaviour

ever_smoked

birthorder <- birthorder %>% mutate(outcome = ever_smoked)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.03 0.0503 -20.47 -1.128 -0.9312 fixed
poly(age, 3, raw = TRUE)1 0.08253 0.004782 17.26 0.07316 0.0919 fixed
poly(age, 3, raw = TRUE)2 -0.001978 0.0001393 -14.2 -0.002251 -0.001705 fixed
poly(age, 3, raw = TRUE)3 0.00001503 0.000001267 11.86 0.00001255 0.00001751 fixed
male 0.6583 0.005408 121.7 0.6477 0.6689 fixed
sibling_count3 -0.01002 0.01094 -0.9159 -0.03148 0.01143 fixed
sibling_count4 -0.001492 0.01126 -0.1325 -0.02357 0.02058 fixed
sibling_count5 0.01039 0.01177 0.8824 -0.01268 0.03345 fixed
sibling_count5+ 0.01886 0.009216 2.047 0.0008013 0.03693 fixed
sd_(Intercept).mother_pidlink 0.1122 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.311 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.03 0.0503 -20.48 -1.129 -0.9315 fixed
birth_order -0.0009195 0.001144 -0.8039 -0.003161 0.001322 fixed
poly(age, 3, raw = TRUE)1 0.0828 0.004794 17.27 0.0734 0.09219 fixed
poly(age, 3, raw = TRUE)2 -0.001988 0.0001398 -14.22 -0.002262 -0.001714 fixed
poly(age, 3, raw = TRUE)3 0.00001512 0.000001272 11.88 0.00001263 0.00001761 fixed
male 0.6583 0.005408 121.7 0.6477 0.6689 fixed
sibling_count3 -0.009801 0.01095 -0.8952 -0.03126 0.01166 fixed
sibling_count4 -0.0008537 0.01129 -0.07562 -0.02298 0.02127 fixed
sibling_count5 0.01149 0.01185 0.9696 -0.01173 0.03471 fixed
sibling_count5+ 0.02234 0.01018 2.195 0.002388 0.04229 fixed
sd_(Intercept).mother_pidlink 0.1122 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.311 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.035 0.05044 -20.52 -1.134 -0.9364 fixed
poly(age, 3, raw = TRUE)1 0.08307 0.004796 17.32 0.07367 0.09247 fixed
poly(age, 3, raw = TRUE)2 -0.001998 0.0001399 -14.29 -0.002272 -0.001724 fixed
poly(age, 3, raw = TRUE)3 0.00001522 0.000001273 11.96 0.00001273 0.00001772 fixed
male 0.6584 0.005408 121.7 0.6478 0.669 fixed
sibling_count3 -0.008598 0.0111 -0.7743 -0.03036 0.01317 fixed
sibling_count4 -0.000003608 0.01161 -0.0003108 -0.02276 0.02275 fixed
sibling_count5 0.01432 0.0123 1.164 -0.009785 0.03844 fixed
sibling_count5+ 0.02736 0.01069 2.558 0.006396 0.04832 fixed
birth_order_nonlinear2 0.006211 0.007852 0.791 -0.009178 0.0216 fixed
birth_order_nonlinear3 -0.005846 0.009241 -0.6326 -0.02396 0.01227 fixed
birth_order_nonlinear4 0.0003569 0.01056 0.0338 -0.02034 0.02105 fixed
birth_order_nonlinear5 -0.01352 0.01202 -1.125 -0.03708 0.01004 fixed
birth_order_nonlinear5+ -0.01148 0.009959 -1.153 -0.031 0.008039 fixed
sd_(Intercept).mother_pidlink 0.1122 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.311 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.041 0.05065 -20.55 -1.14 -0.9414 fixed
poly(age, 3, raw = TRUE)1 0.08302 0.004795 17.31 0.07362 0.09242 fixed
poly(age, 3, raw = TRUE)2 -0.002001 0.0001399 -14.31 -0.002275 -0.001727 fixed
poly(age, 3, raw = TRUE)3 0.00001528 0.000001273 12.01 0.00001279 0.00001778 fixed
male 0.6581 0.005408 121.7 0.6475 0.6687 fixed
count_birth_order2/2 0.02629 0.01527 1.722 -0.003631 0.05621 fixed
count_birth_order1/3 -0.01405 0.0145 -0.9691 -0.04248 0.01437 fixed
count_birth_order2/3 0.01445 0.01621 0.8916 -0.01732 0.04623 fixed
count_birth_order3/3 0.00651 0.01806 0.3604 -0.0289 0.04192 fixed
count_birth_order1/4 0.01338 0.01641 0.8149 -0.01879 0.04555 fixed
count_birth_order2/4 0.006355 0.01745 0.3641 -0.02785 0.04056 fixed
count_birth_order3/4 -0.001847 0.01893 -0.09757 -0.03895 0.03526 fixed
count_birth_order4/4 0.01231 0.01984 0.6205 -0.02658 0.0512 fixed
count_birth_order1/5 0.03854 0.01879 2.051 0.001707 0.07537 fixed
count_birth_order2/5 0.005322 0.01985 0.2681 -0.03359 0.04423 fixed
count_birth_order3/5 0.003385 0.02026 0.167 -0.03633 0.0431 fixed
count_birth_order4/5 0.05807 0.0217 2.676 0.01554 0.1006 fixed
count_birth_order5/5 -0.008625 0.02211 -0.3901 -0.05196 0.03471 fixed
count_birth_order1/5+ 0.04767 0.01512 3.152 0.01803 0.0773 fixed
count_birth_order2/5+ 0.03767 0.01564 2.409 0.007016 0.06833 fixed
count_birth_order3/5+ 0.02911 0.01529 1.903 -0.0008641 0.05908 fixed
count_birth_order4/5+ 0.0213 0.01504 1.417 -0.008173 0.05078 fixed
count_birth_order5/5+ 0.02724 0.01511 1.802 -0.00238 0.05687 fixed
count_birth_order5+/5+ 0.02345 0.01185 1.979 0.0002202 0.04667 fixed
sd_(Intercept).mother_pidlink 0.1121 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.311 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 9037 9120 -4507 9015 NA NA NA
12 9038 9129 -4507 9014 0.6469 1 0.4212
16 9042 9164 -4505 9010 3.709 4 0.4468
26 9046 9244 -4497 8994 16.27 10 0.09213

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.811 0.1473 -12.3 -2.1 -1.523 fixed
poly(age, 3, raw = TRUE)1 0.175 0.01672 10.47 0.1423 0.2078 fixed
poly(age, 3, raw = TRUE)2 -0.005302 0.0005975 -8.873 -0.006473 -0.004131 fixed
poly(age, 3, raw = TRUE)3 0.00005222 0.000006778 7.704 0.00003893 0.0000655 fixed
male 0.5873 0.008561 68.6 0.5705 0.6041 fixed
sibling_count3 0.007451 0.0136 0.5479 -0.0192 0.03411 fixed
sibling_count4 0.004702 0.01467 0.3206 -0.02405 0.03345 fixed
sibling_count5 0.03346 0.01688 1.982 0.0003739 0.06655 fixed
sibling_count5+ 0.0422 0.01476 2.859 0.01327 0.07113 fixed
sd_(Intercept).mother_pidlink 0.135 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3135 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.811 0.1473 -12.29 -2.1 -1.522 fixed
birth_order -0.0006785 0.002855 -0.2377 -0.006274 0.004917 fixed
poly(age, 3, raw = TRUE)1 0.1751 0.01672 10.47 0.1423 0.2078 fixed
poly(age, 3, raw = TRUE)2 -0.005302 0.0005976 -8.872 -0.006473 -0.004131 fixed
poly(age, 3, raw = TRUE)3 0.00005219 0.000006779 7.698 0.0000389 0.00006548 fixed
male 0.5873 0.008562 68.59 0.5705 0.6041 fixed
sibling_count3 0.007791 0.01368 0.5697 -0.01901 0.0346 fixed
sibling_count4 0.005488 0.01504 0.365 -0.02398 0.03496 fixed
sibling_count5 0.03474 0.01772 1.96 0.000007231 0.06947 fixed
sibling_count5+ 0.04478 0.01832 2.445 0.008878 0.08068 fixed
sd_(Intercept).mother_pidlink 0.135 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3135 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.823 0.1476 -12.35 -2.113 -1.534 fixed
poly(age, 3, raw = TRUE)1 0.1759 0.01674 10.5 0.1431 0.2087 fixed
poly(age, 3, raw = TRUE)2 -0.005327 0.0005982 -8.905 -0.0065 -0.004155 fixed
poly(age, 3, raw = TRUE)3 0.00005238 0.000006787 7.718 0.00003908 0.00006568 fixed
male 0.5875 0.008564 68.6 0.5707 0.6043 fixed
sibling_count3 0.006373 0.01395 0.4567 -0.02098 0.03372 fixed
sibling_count4 0.005122 0.01561 0.3282 -0.02547 0.03571 fixed
sibling_count5 0.03905 0.01858 2.102 0.002639 0.07546 fixed
sibling_count5+ 0.05275 0.01885 2.798 0.01579 0.0897 fixed
birth_order_nonlinear2 0.01327 0.01101 1.205 -0.008314 0.03486 fixed
birth_order_nonlinear3 0.005947 0.01359 0.4377 -0.02069 0.03258 fixed
birth_order_nonlinear4 -0.002039 0.01687 -0.1209 -0.0351 0.03103 fixed
birth_order_nonlinear5 -0.0184 0.02114 -0.8704 -0.05983 0.02303 fixed
birth_order_nonlinear5+ -0.01167 0.02127 -0.5488 -0.05335 0.03001 fixed
sd_(Intercept).mother_pidlink 0.1349 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3136 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.824 0.1479 -12.33 -2.114 -1.534 fixed
poly(age, 3, raw = TRUE)1 0.175 0.01678 10.43 0.1421 0.2079 fixed
poly(age, 3, raw = TRUE)2 -0.005298 0.0005996 -8.836 -0.006473 -0.004123 fixed
poly(age, 3, raw = TRUE)3 0.00005206 0.000006804 7.651 0.00003872 0.0000654 fixed
male 0.5874 0.008572 68.52 0.5706 0.6042 fixed
count_birth_order2/2 0.03734 0.01984 1.882 -0.001552 0.07623 fixed
count_birth_order1/3 0.02279 0.01783 1.278 -0.01215 0.05772 fixed
count_birth_order2/3 0.01879 0.01937 0.9702 -0.01917 0.05676 fixed
count_birth_order3/3 0.01655 0.02162 0.7655 -0.02582 0.05892 fixed
count_birth_order1/4 0.006185 0.0217 0.285 -0.03634 0.04871 fixed
count_birth_order2/4 0.0236 0.02242 1.053 -0.02033 0.06754 fixed
count_birth_order3/4 0.03506 0.02373 1.477 -0.01146 0.08157 fixed
count_birth_order4/4 0.007861 0.02467 0.3187 -0.04048 0.0562 fixed
count_birth_order1/5 0.03545 0.0295 1.202 -0.02237 0.09326 fixed
count_birth_order2/5 0.06082 0.0317 1.918 -0.00132 0.123 fixed
count_birth_order3/5 0.05973 0.02949 2.026 0.001939 0.1175 fixed
count_birth_order4/5 0.04242 0.02892 1.467 -0.01426 0.09909 fixed
count_birth_order5/5 0.03742 0.02999 1.248 -0.02136 0.0962 fixed
count_birth_order1/5+ 0.09227 0.0293 3.149 0.03485 0.1497 fixed
count_birth_order2/5+ 0.06653 0.02937 2.265 0.008968 0.1241 fixed
count_birth_order3/5+ 0.04338 0.02873 1.51 -0.01293 0.09969 fixed
count_birth_order4/5+ 0.06546 0.02705 2.42 0.01245 0.1185 fixed
count_birth_order5/5+ 0.03659 0.02583 1.417 -0.01403 0.08721 fixed
count_birth_order5+/5+ 0.04926 0.01961 2.512 0.01083 0.0877 fixed
sd_(Intercept).mother_pidlink 0.1344 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3139 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 4222 4296 -2100 4200 NA NA NA
12 4224 4304 -2100 4200 0.05593 1 0.813
16 4228 4336 -2098 4196 3.26 4 0.5153
26 4242 4417 -2095 4190 6.288 10 0.7905

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.794 0.1469 -12.21 -2.081 -1.506 fixed
poly(age, 3, raw = TRUE)1 0.1732 0.01669 10.38 0.1405 0.2059 fixed
poly(age, 3, raw = TRUE)2 -0.005237 0.0005964 -8.78 -0.006406 -0.004068 fixed
poly(age, 3, raw = TRUE)3 0.00005152 0.000006767 7.614 0.00003826 0.00006479 fixed
male 0.5869 0.008534 68.77 0.5702 0.6036 fixed
sibling_count3 0.008936 0.01474 0.6065 -0.01994 0.03782 fixed
sibling_count4 -0.001759 0.01552 -0.1134 -0.03218 0.02866 fixed
sibling_count5 0.02235 0.0167 1.338 -0.01038 0.05507 fixed
sibling_count5+ 0.0372 0.01461 2.547 0.008571 0.06583 fixed
sd_(Intercept).mother_pidlink 0.1351 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3138 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.794 0.1469 -12.21 -2.082 -1.506 fixed
birth_order 0.00006028 0.002496 0.02415 -0.004833 0.004953 fixed
poly(age, 3, raw = TRUE)1 0.1732 0.01669 10.38 0.1405 0.2059 fixed
poly(age, 3, raw = TRUE)2 -0.005237 0.0005965 -8.779 -0.006406 -0.004068 fixed
poly(age, 3, raw = TRUE)3 0.00005153 0.000006769 7.612 0.00003826 0.00006479 fixed
male 0.5869 0.008535 68.77 0.5702 0.6036 fixed
sibling_count3 0.008906 0.01479 0.6023 -0.02008 0.03789 fixed
sibling_count4 -0.001826 0.01577 -0.1158 -0.03273 0.02907 fixed
sibling_count5 0.02224 0.01726 1.289 -0.01158 0.05606 fixed
sibling_count5+ 0.03698 0.01725 2.144 0.003169 0.07079 fixed
sd_(Intercept).mother_pidlink 0.1351 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3138 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.805 0.1472 -12.26 -2.094 -1.517 fixed
poly(age, 3, raw = TRUE)1 0.1739 0.0167 10.41 0.1412 0.2067 fixed
poly(age, 3, raw = TRUE)2 -0.005261 0.000597 -8.812 -0.006431 -0.004091 fixed
poly(age, 3, raw = TRUE)3 0.00005171 0.000006775 7.631 0.00003843 0.00006499 fixed
male 0.5871 0.008536 68.78 0.5704 0.6038 fixed
sibling_count3 0.009283 0.01506 0.6165 -0.02023 0.0388 fixed
sibling_count4 -0.001904 0.01631 -0.1168 -0.03386 0.03005 fixed
sibling_count5 0.02286 0.01805 1.267 -0.01251 0.05823 fixed
sibling_count5+ 0.04524 0.01778 2.545 0.0104 0.08008 fixed
birth_order_nonlinear2 0.01394 0.01123 1.241 -0.008069 0.03595 fixed
birth_order_nonlinear3 -0.0001939 0.01354 -0.01432 -0.02674 0.02635 fixed
birth_order_nonlinear4 0.006623 0.01636 0.4047 -0.02545 0.0387 fixed
birth_order_nonlinear5 0.004426 0.02007 0.2205 -0.03492 0.04377 fixed
birth_order_nonlinear5+ -0.01611 0.01903 -0.8463 -0.05342 0.0212 fixed
sd_(Intercept).mother_pidlink 0.1356 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3136 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.817 0.1475 -12.31 -2.106 -1.528 fixed
poly(age, 3, raw = TRUE)1 0.1742 0.01673 10.41 0.1414 0.207 fixed
poly(age, 3, raw = TRUE)2 -0.005268 0.0005982 -8.807 -0.00644 -0.004096 fixed
poly(age, 3, raw = TRUE)3 0.00005175 0.000006789 7.623 0.00003845 0.00006506 fixed
male 0.5867 0.008543 68.68 0.57 0.6035 fixed
count_birth_order2/2 0.04181 0.0218 1.918 -0.0009097 0.08454 fixed
count_birth_order1/3 0.0185 0.01933 0.9574 -0.01937 0.05638 fixed
count_birth_order2/3 0.0225 0.02097 1.073 -0.01859 0.0636 fixed
count_birth_order3/3 0.03344 0.02342 1.428 -0.01246 0.07935 fixed
count_birth_order1/4 0.0112 0.0227 0.4936 -0.03328 0.05569 fixed
count_birth_order2/4 0.03426 0.02315 1.48 -0.0111 0.07963 fixed
count_birth_order3/4 -0.0004727 0.02537 -0.01863 -0.05019 0.04925 fixed
count_birth_order4/4 -0.00158 0.02624 -0.06022 -0.05301 0.04985 fixed
count_birth_order1/5 0.05073 0.02702 1.877 -0.00223 0.1037 fixed
count_birth_order2/5 0.03313 0.02898 1.143 -0.02368 0.08994 fixed
count_birth_order3/5 0.0241 0.02834 0.8502 -0.03145 0.07964 fixed
count_birth_order4/5 0.03857 0.02958 1.304 -0.01941 0.09654 fixed
count_birth_order5/5 0.03566 0.02946 1.211 -0.02207 0.09339 fixed
count_birth_order1/5+ 0.06043 0.02573 2.348 0.009995 0.1109 fixed
count_birth_order2/5+ 0.05178 0.02694 1.922 -0.001035 0.1046 fixed
count_birth_order3/5+ 0.04945 0.02611 1.894 -0.001733 0.1006 fixed
count_birth_order4/5+ 0.07459 0.02516 2.965 0.02528 0.1239 fixed
count_birth_order5/5+ 0.05976 0.02593 2.305 0.00894 0.1106 fixed
count_birth_order5+/5+ 0.03856 0.01928 2.001 0.0007835 0.07634 fixed
sd_(Intercept).mother_pidlink 0.1352 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3139 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 4271 4345 -2124 4249 NA NA NA
12 4273 4354 -2124 4249 0.0006262 1 0.98
16 4277 4385 -2123 4245 3.427 4 0.489
26 4291 4467 -2120 4239 6.031 10 0.8126

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.855 0.1489 -12.46 -2.147 -1.564 fixed
poly(age, 3, raw = TRUE)1 0.1807 0.01692 10.68 0.1475 0.2139 fixed
poly(age, 3, raw = TRUE)2 -0.005507 0.0006047 -9.107 -0.006692 -0.004322 fixed
poly(age, 3, raw = TRUE)3 0.0000546 0.000006863 7.955 0.00004115 0.00006805 fixed
male 0.5873 0.008645 67.94 0.5704 0.6042 fixed
sibling_count3 0.002991 0.01342 0.2229 -0.02331 0.02929 fixed
sibling_count4 -0.001225 0.01458 -0.08399 -0.0298 0.02735 fixed
sibling_count5 0.01576 0.01729 0.9113 -0.01813 0.04964 fixed
sibling_count5+ 0.03472 0.01492 2.327 0.00548 0.06395 fixed
sd_(Intercept).mother_pidlink 0.1349 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3134 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.855 0.1489 -12.46 -2.147 -1.563 fixed
birth_order -0.000005795 0.002939 -0.001972 -0.005766 0.005754 fixed
poly(age, 3, raw = TRUE)1 0.1807 0.01692 10.68 0.1475 0.2139 fixed
poly(age, 3, raw = TRUE)2 -0.005507 0.0006048 -9.106 -0.006692 -0.004322 fixed
poly(age, 3, raw = TRUE)3 0.0000546 0.000006865 7.954 0.00004114 0.00006805 fixed
male 0.5873 0.008646 67.93 0.5704 0.6042 fixed
sibling_count3 0.002994 0.0135 0.2218 -0.02347 0.02946 fixed
sibling_count4 -0.001218 0.01496 -0.08141 -0.03054 0.0281 fixed
sibling_count5 0.01577 0.01809 0.8717 -0.01968 0.05122 fixed
sibling_count5+ 0.03474 0.01857 1.871 -0.00165 0.07113 fixed
sd_(Intercept).mother_pidlink 0.1349 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3134 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.869 0.1492 -12.53 -2.162 -1.577 fixed
poly(age, 3, raw = TRUE)1 0.1817 0.01694 10.73 0.1485 0.2149 fixed
poly(age, 3, raw = TRUE)2 -0.005537 0.0006053 -9.147 -0.006723 -0.004351 fixed
poly(age, 3, raw = TRUE)3 0.00005483 0.000006871 7.979 0.00004136 0.00006829 fixed
male 0.5875 0.008647 67.94 0.5705 0.6044 fixed
sibling_count3 0.001975 0.01379 0.1433 -0.02505 0.029 fixed
sibling_count4 -0.0008761 0.01555 -0.05635 -0.03135 0.0296 fixed
sibling_count5 0.02003 0.01889 1.06 -0.017 0.05706 fixed
sibling_count5+ 0.04395 0.01914 2.296 0.006439 0.08146 fixed
birth_order_nonlinear2 0.01494 0.01099 1.36 -0.006592 0.03647 fixed
birth_order_nonlinear3 0.006761 0.01359 0.4976 -0.01987 0.03339 fixed
birth_order_nonlinear4 -0.002179 0.01735 -0.1256 -0.03619 0.03183 fixed
birth_order_nonlinear5 -0.01151 0.022 -0.5233 -0.05462 0.0316 fixed
birth_order_nonlinear5+ -0.01037 0.02192 -0.4732 -0.05333 0.03259 fixed
sd_(Intercept).mother_pidlink 0.135 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3134 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -1.866 0.1496 -12.47 -2.159 -1.573 fixed
poly(age, 3, raw = TRUE)1 0.1803 0.01697 10.62 0.147 0.2136 fixed
poly(age, 3, raw = TRUE)2 -0.005484 0.0006068 -9.037 -0.006673 -0.004294 fixed
poly(age, 3, raw = TRUE)3 0.00005418 0.00000689 7.864 0.00004068 0.00006769 fixed
male 0.5871 0.008655 67.84 0.5702 0.6041 fixed
count_birth_order2/2 0.03855 0.01928 2 0.0007658 0.07634 fixed
count_birth_order1/3 0.01712 0.0176 0.973 -0.01737 0.05162 fixed
count_birth_order2/3 0.01647 0.01936 0.8508 -0.02147 0.05442 fixed
count_birth_order3/3 0.01444 0.02124 0.6799 -0.02719 0.05608 fixed
count_birth_order1/4 -0.006943 0.02179 -0.3187 -0.04964 0.03576 fixed
count_birth_order2/4 0.02816 0.02244 1.255 -0.01583 0.07215 fixed
count_birth_order3/4 0.02373 0.02352 1.009 -0.02236 0.06982 fixed
count_birth_order4/4 0.006673 0.02483 0.2687 -0.042 0.05534 fixed
count_birth_order1/5 0.03185 0.0295 1.08 -0.02596 0.08967 fixed
count_birth_order2/5 0.02865 0.03271 0.876 -0.03545 0.09276 fixed
count_birth_order3/5 0.03949 0.03081 1.282 -0.02089 0.09986 fixed
count_birth_order4/5 0.01282 0.03022 0.4241 -0.04641 0.07204 fixed
count_birth_order5/5 0.03525 0.03195 1.103 -0.02737 0.09787 fixed
count_birth_order1/5+ 0.09172 0.03002 3.055 0.03288 0.1506 fixed
count_birth_order2/5+ 0.04846 0.0301 1.61 -0.01054 0.1075 fixed
count_birth_order3/5+ 0.04365 0.02902 1.504 -0.01323 0.1005 fixed
count_birth_order4/5+ 0.05895 0.02846 2.071 0.003172 0.1147 fixed
count_birth_order5/5+ 0.02874 0.0264 1.089 -0.023 0.08049 fixed
count_birth_order5+/5+ 0.04156 0.01996 2.082 0.002429 0.08069 fixed
sd_(Intercept).mother_pidlink 0.1344 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.3137 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 4133 4207 -2056 4111 NA NA NA
12 4135 4216 -2056 4111 0.0000002846 1 0.9996
16 4140 4248 -2054 4108 3.094 4 0.5423
26 4153 4327 -2050 4101 7.745 10 0.6537

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

still_smoking

birthorder <- birthorder %>% mutate(outcome = still_smoking)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)

Naive birth order

outcome_naive_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_naive_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.4994 0.07927 6.299 0.344 0.6547 fixed
poly(age, 3, raw = TRUE)1 0.02068 0.006708 3.082 0.00753 0.03383 fixed
poly(age, 3, raw = TRUE)2 -0.0005941 0.0001832 -3.243 -0.0009531 -0.0002351 fixed
poly(age, 3, raw = TRUE)3 0.000004817 0.000001566 3.076 0.000001748 0.000007886 fixed
male 0.2089 0.02436 8.575 0.1611 0.2566 fixed
sibling_count3 0.01578 0.01612 0.9789 -0.01581 0.04737 fixed
sibling_count4 0.003225 0.01616 0.1996 -0.02844 0.03489 fixed
sibling_count5 0.009782 0.01671 0.5853 -0.02297 0.04253 fixed
sibling_count5+ 0.0114 0.01314 0.8676 -0.01435 0.03715 fixed
sd_(Intercept).mother_pidlink 0.03882 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2719 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.4993 0.07928 6.297 0.3439 0.6547 fixed
birth_order -0.0001987 0.001539 -0.1291 -0.003215 0.002818 fixed
poly(age, 3, raw = TRUE)1 0.02074 0.006724 3.084 0.007557 0.03391 fixed
poly(age, 3, raw = TRUE)2 -0.000596 0.0001838 -3.243 -0.0009562 -0.0002358 fixed
poly(age, 3, raw = TRUE)3 0.000004833 0.000001571 3.076 0.000001754 0.000007912 fixed
male 0.2089 0.02436 8.575 0.1612 0.2566 fixed
sibling_count3 0.01584 0.01613 0.9821 -0.01577 0.04745 fixed
sibling_count4 0.003347 0.01619 0.2068 -0.02838 0.03507 fixed
sibling_count5 0.009997 0.0168 0.5952 -0.02292 0.04292 fixed
sibling_count5+ 0.01211 0.01424 0.85 -0.01581 0.04002 fixed
sd_(Intercept).mother_pidlink 0.03895 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2719 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.4961 0.07939 6.249 0.3405 0.6517 fixed
poly(age, 3, raw = TRUE)1 0.02088 0.006727 3.104 0.007696 0.03406 fixed
poly(age, 3, raw = TRUE)2 -0.0005984 0.0001838 -3.256 -0.0009587 -0.0002382 fixed
poly(age, 3, raw = TRUE)3 0.000004839 0.000001571 3.081 0.000001761 0.000007918 fixed
male 0.2098 0.02437 8.608 0.162 0.2575 fixed
sibling_count3 0.02091 0.01643 1.272 -0.0113 0.05311 fixed
sibling_count4 0.005396 0.01667 0.3237 -0.02728 0.03807 fixed
sibling_count5 0.01371 0.01745 0.7858 -0.02049 0.04792 fixed
sibling_count5+ 0.01531 0.01499 1.021 -0.01407 0.0447 fixed
birth_order_nonlinear2 0.00008096 0.01166 0.006942 -0.02278 0.02294 fixed
birth_order_nonlinear3 -0.01968 0.01347 -1.462 -0.04608 0.006712 fixed
birth_order_nonlinear4 0.009552 0.01505 0.6349 -0.01994 0.03904 fixed
birth_order_nonlinear5 -0.01544 0.01716 -0.8998 -0.04908 0.01819 fixed
birth_order_nonlinear5+ -0.002691 0.01378 -0.1952 -0.02971 0.02432 fixed
sd_(Intercept).mother_pidlink 0.03821 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.272 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) 0.493 0.07968 6.187 0.3368 0.6491 fixed
poly(age, 3, raw = TRUE)1 0.02063 0.006731 3.065 0.007438 0.03382 fixed
poly(age, 3, raw = TRUE)2 -0.0005914 0.0001838 -3.217 -0.0009517 -0.0002311 fixed
poly(age, 3, raw = TRUE)3 0.000004788 0.00000157 3.049 0.00000171 0.000007866 fixed
male 0.2092 0.02437 8.584 0.1614 0.2569 fixed
count_birth_order2/2 0.01447 0.02335 0.6196 -0.0313 0.06024 fixed
count_birth_order1/3 0.02399 0.02313 1.037 -0.02135 0.06933 fixed
count_birth_order2/3 0.03065 0.0246 1.246 -0.01756 0.07886 fixed
count_birth_order3/3 0.007198 0.02658 0.2708 -0.04489 0.05929 fixed
count_birth_order1/4 0.0006132 0.0246 0.02493 -0.04759 0.04882 fixed
count_birth_order2/4 0.008111 0.02642 0.307 -0.04368 0.0599 fixed
count_birth_order3/4 -0.001621 0.02848 -0.05692 -0.05744 0.0542 fixed
count_birth_order4/4 0.03651 0.02921 1.25 -0.02074 0.09375 fixed
count_birth_order1/5 0.04653 0.0274 1.698 -0.007168 0.1002 fixed
count_birth_order2/5 -0.02254 0.03016 -0.7475 -0.08165 0.03656 fixed
count_birth_order3/5 -0.04164 0.03144 -1.324 -0.1033 0.01998 fixed
count_birth_order4/5 0.02983 0.03076 0.9699 -0.03045 0.09012 fixed
count_birth_order5/5 0.06484 0.03397 1.908 -0.001749 0.1314 fixed
count_birth_order1/5+ 0.02421 0.02135 1.134 -0.01763 0.06605 fixed
count_birth_order2/5+ 0.03074 0.02247 1.368 -0.0133 0.07477 fixed
count_birth_order3/5+ 0.01257 0.02201 0.5714 -0.03056 0.0557 fixed
count_birth_order4/5+ 0.02479 0.02202 1.126 -0.01837 0.06795 fixed
count_birth_order5/5+ -0.01014 0.02209 -0.4591 -0.05344 0.03316 fixed
count_birth_order5+/5+ 0.01887 0.01779 1.06 -0.016 0.05374 fixed
sd_(Intercept).mother_pidlink 0.03823 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2719 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_naive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 1276 1348 -627.1 1254 NA NA NA
12 1278 1356 -627.1 1254 0.01671 1 0.8972
16 1282 1386 -624.8 1250 4.631 4 0.3273
26 1288 1457 -617.9 1236 13.77 10 0.1836

Maternal birth order

outcome_uterus_m1 <- update(m2_birthorder_linear, data = birthorder %>%
                     mutate(sibling_count = sibling_count_uterus_alive_factor,
                            birth_order_nonlinear = birthorder_uterus_alive_factor,
                            birth_order = birthorder_uterus_alive,
                            count_birth_order = count_birthorder_uterus_alive) %>% 
                     filter(sibling_count != "1"))
compare_models_markdown(outcome_uterus_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.08365 0.2385 -0.3507 -0.5511 0.3838 fixed
poly(age, 3, raw = TRUE)1 0.08334 0.02526 3.299 0.03383 0.1328 fixed
poly(age, 3, raw = TRUE)2 -0.002847 0.0008678 -3.281 -0.004548 -0.001146 fixed
poly(age, 3, raw = TRUE)3 0.00003 0.000009506 3.156 0.00001136 0.00004863 fixed
male 0.2353 0.04151 5.669 0.154 0.3167 fixed
sibling_count3 0.008865 0.02092 0.4238 -0.03213 0.04986 fixed
sibling_count4 0.03489 0.02196 1.589 -0.008154 0.07793 fixed
sibling_count5 0.02659 0.02399 1.108 -0.02043 0.0736 fixed
sibling_count5+ 0.03027 0.02075 1.459 -0.01041 0.07094 fixed
sd_(Intercept).mother_pidlink 0 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2791 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.08354 0.2386 -0.3502 -0.5511 0.384 fixed
birth_order -0.001376 0.003918 -0.3513 -0.009056 0.006303 fixed
poly(age, 3, raw = TRUE)1 0.08351 0.02527 3.305 0.03398 0.133 fixed
poly(age, 3, raw = TRUE)2 -0.00285 0.0008681 -3.283 -0.004551 -0.001149 fixed
poly(age, 3, raw = TRUE)3 0.00002997 0.000009508 3.152 0.00001133 0.00004861 fixed
male 0.2352 0.04153 5.664 0.1538 0.3166 fixed
sibling_count3 0.009502 0.021 0.4525 -0.03166 0.05066 fixed
sibling_count4 0.03644 0.0224 1.626 -0.007474 0.08035 fixed
sibling_count5 0.02904 0.02499 1.162 -0.01993 0.07801 fixed
sibling_count5+ 0.03525 0.02514 1.402 -0.01402 0.08452 fixed
sd_(Intercept).mother_pidlink 0.003056 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2792 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.08294 0.2389 -0.3471 -0.5513 0.3854 fixed
poly(age, 3, raw = TRUE)1 0.08364 0.02531 3.305 0.03404 0.1332 fixed
poly(age, 3, raw = TRUE)2 -0.002862 0.0008694 -3.292 -0.004566 -0.001158 fixed
poly(age, 3, raw = TRUE)3 0.00003026 0.000009525 3.177 0.00001159 0.00004893 fixed
male 0.2362 0.04156 5.683 0.1547 0.3176 fixed
sibling_count3 0.01247 0.02145 0.5814 -0.02958 0.05452 fixed
sibling_count4 0.04025 0.02337 1.722 -0.00556 0.08606 fixed
sibling_count5 0.0282 0.02649 1.064 -0.02372 0.08012 fixed
sibling_count5+ 0.02489 0.02601 0.9569 -0.02609 0.07587 fixed
birth_order_nonlinear2 -0.009683 0.01749 -0.5536 -0.04396 0.0246 fixed
birth_order_nonlinear3 -0.01878 0.02057 -0.9128 -0.05909 0.02154 fixed
birth_order_nonlinear4 -0.01068 0.02504 -0.4265 -0.05976 0.0384 fixed
birth_order_nonlinear5 0.01235 0.03151 0.3919 -0.0494 0.0741 fixed
birth_order_nonlinear5+ 0.01187 0.02963 0.4004 -0.04622 0.06995 fixed
sd_(Intercept).mother_pidlink 0 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2793 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.098 0.2399 -0.4085 -0.5682 0.3722 fixed
poly(age, 3, raw = TRUE)1 0.08341 0.02541 3.282 0.0336 0.1332 fixed
poly(age, 3, raw = TRUE)2 -0.002855 0.0008731 -3.27 -0.004566 -0.001144 fixed
poly(age, 3, raw = TRUE)3 0.00003018 0.000009567 3.155 0.00001143 0.00004894 fixed
male 0.2385 0.04161 5.731 0.1569 0.32 fixed
count_birth_order2/2 0.02927 0.03273 0.8945 -0.03487 0.09341 fixed
count_birth_order1/3 0.01775 0.02863 0.6201 -0.03836 0.07387 fixed
count_birth_order2/3 0.0114 0.03194 0.3568 -0.05121 0.074 fixed
count_birth_order3/3 0.03532 0.03424 1.031 -0.0318 0.1024 fixed
count_birth_order1/4 0.09363 0.03488 2.684 0.02527 0.162 fixed
count_birth_order2/4 0.0124 0.03513 0.3529 -0.05646 0.08126 fixed
count_birth_order3/4 0.029 0.03633 0.7981 -0.04221 0.1002 fixed
count_birth_order4/4 0.04567 0.03891 1.174 -0.0306 0.1219 fixed
count_birth_order1/5 0.06573 0.04413 1.49 -0.02075 0.1522 fixed
count_birth_order2/5 0.0229 0.04788 0.4783 -0.07094 0.1167 fixed
count_birth_order3/5 0.005765 0.04394 0.1312 -0.08035 0.09188 fixed
count_birth_order4/5 0.06116 0.04205 1.454 -0.02127 0.1436 fixed
count_birth_order5/5 0.02552 0.04559 0.5598 -0.06383 0.1149 fixed
count_birth_order1/5+ 0.03691 0.03967 0.9304 -0.04085 0.1147 fixed
count_birth_order2/5+ 0.05884 0.04275 1.376 -0.02494 0.1426 fixed
count_birth_order3/5+ 0.004336 0.0406 0.1068 -0.07523 0.0839 fixed
count_birth_order4/5+ 0.005484 0.03915 0.1401 -0.07126 0.08223 fixed
count_birth_order5/5+ 0.07248 0.03947 1.836 -0.004882 0.1498 fixed
count_birth_order5+/5+ 0.05185 0.02905 1.785 -0.005084 0.1088 fixed
sd_(Intercept).mother_pidlink 0.0000002143 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2794 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_alive_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 546.4 607.3 -262.2 524.4 NA NA NA
12 548.3 614.7 -262.2 524.3 0.124 1 0.7248
16 554.5 643.1 -261.3 522.5 1.784 4 0.7754
26 564.9 708.8 -256.5 512.9 9.584 10 0.4777

Maternal pregnancy order

outcome_preg_m1 <- update(m2_birthorder_linear, data = birthorder %>% 
  mutate(sibling_count = sibling_count_uterus_preg_factor,
         birth_order_nonlinear = birthorder_uterus_preg_factor,
         birth_order = birthorder_uterus_preg,
         count_birth_order = count_birthorder_uterus_preg
         ) %>% 
    filter(sibling_count != "1"))
compare_models_markdown(outcome_preg_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.06947 0.2387 -0.291 -0.5374 0.3984 fixed
poly(age, 3, raw = TRUE)1 0.08058 0.02532 3.182 0.03094 0.1302 fixed
poly(age, 3, raw = TRUE)2 -0.002747 0.0008703 -3.156 -0.004453 -0.001041 fixed
poly(age, 3, raw = TRUE)3 0.00002892 0.000009536 3.032 0.00001023 0.00004761 fixed
male 0.2456 0.04089 6.007 0.1655 0.3258 fixed
sibling_count3 0.02061 0.02307 0.8931 -0.02461 0.06582 fixed
sibling_count4 0.01468 0.02371 0.619 -0.03179 0.06114 fixed
sibling_count5 0.001525 0.02502 0.06093 -0.04751 0.05056 fixed
sibling_count5+ 0.03133 0.02169 1.444 -0.01119 0.07385 fixed
sd_(Intercept).mother_pidlink 0 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2809 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.06945 0.2388 -0.2908 -0.5375 0.3986 fixed
birth_order 0.0005249 0.003441 0.1525 -0.00622 0.007269 fixed
poly(age, 3, raw = TRUE)1 0.0805 0.02533 3.177 0.03085 0.1302 fixed
poly(age, 3, raw = TRUE)2 -0.002745 0.0008706 -3.153 -0.004452 -0.001039 fixed
poly(age, 3, raw = TRUE)3 0.00002892 0.000009539 3.032 0.00001023 0.00004762 fixed
male 0.2457 0.0409 6.006 0.1655 0.3258 fixed
sibling_count3 0.02035 0.02314 0.8797 -0.02499 0.0657 fixed
sibling_count4 0.01415 0.02397 0.5904 -0.03282 0.06112 fixed
sibling_count5 0.0006978 0.02561 0.02725 -0.04949 0.05089 fixed
sibling_count5+ 0.02945 0.02495 1.18 -0.01946 0.07836 fixed
sd_(Intercept).mother_pidlink 0 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.281 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.06587 0.2393 -0.2753 -0.5349 0.4031 fixed
poly(age, 3, raw = TRUE)1 0.08039 0.02539 3.166 0.03063 0.1301 fixed
poly(age, 3, raw = TRUE)2 -0.002747 0.0008723 -3.149 -0.004457 -0.001037 fixed
poly(age, 3, raw = TRUE)3 0.00002904 0.000009559 3.038 0.00001031 0.00004778 fixed
male 0.2466 0.04095 6.022 0.1663 0.3269 fixed
sibling_count3 0.02183 0.02363 0.9241 -0.02447 0.06814 fixed
sibling_count4 0.01509 0.02475 0.6096 -0.03342 0.06359 fixed
sibling_count5 0.000362 0.02684 0.01349 -0.05225 0.05297 fixed
sibling_count5+ 0.0223 0.02591 0.8605 -0.02848 0.07307 fixed
birth_order_nonlinear2 -0.005389 0.01784 -0.3021 -0.04035 0.02957 fixed
birth_order_nonlinear3 -0.007488 0.0207 -0.3618 -0.04805 0.03308 fixed
birth_order_nonlinear4 0.001033 0.02432 0.04248 -0.04663 0.0487 fixed
birth_order_nonlinear5 0.00444 0.03014 0.1473 -0.05463 0.06351 fixed
birth_order_nonlinear5+ 0.0206 0.02719 0.7578 -0.03269 0.0739 fixed
sd_(Intercept).mother_pidlink 0.0000001225 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2812 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.0715 0.2401 -0.2978 -0.5421 0.3991 fixed
poly(age, 3, raw = TRUE)1 0.08093 0.02546 3.179 0.03104 0.1308 fixed
poly(age, 3, raw = TRUE)2 -0.002754 0.0008747 -3.148 -0.004468 -0.001039 fixed
poly(age, 3, raw = TRUE)3 0.000029 0.000009585 3.026 0.00001022 0.00004779 fixed
male 0.2473 0.04103 6.026 0.1669 0.3277 fixed
count_birth_order2/2 -0.01316 0.03653 -0.3603 -0.08475 0.05843 fixed
count_birth_order1/3 0.00648 0.03199 0.2026 -0.05622 0.06918 fixed
count_birth_order2/3 0.02276 0.03503 0.6497 -0.0459 0.09142 fixed
count_birth_order3/3 0.02057 0.03696 0.5564 -0.05188 0.09301 fixed
count_birth_order1/4 0.004074 0.03622 0.1125 -0.06692 0.07506 fixed
count_birth_order2/4 -0.0125 0.0367 -0.3405 -0.08443 0.05943 fixed
count_birth_order3/4 0.01958 0.0411 0.4763 -0.06098 0.1001 fixed
count_birth_order4/4 0.03758 0.04186 0.8979 -0.04446 0.1196 fixed
count_birth_order1/5 0.03448 0.0407 0.8472 -0.04529 0.1143 fixed
count_birth_order2/5 -0.06708 0.04668 -1.437 -0.1586 0.02441 fixed
count_birth_order3/5 -0.01311 0.0456 -0.2876 -0.1025 0.07626 fixed
count_birth_order4/5 0.01589 0.04613 0.3446 -0.07451 0.1063 fixed
count_birth_order5/5 -0.01125 0.04681 -0.2404 -0.103 0.08049 fixed
count_birth_order1/5+ 0.01485 0.03818 0.3891 -0.05998 0.08969 fixed
count_birth_order2/5+ 0.0675 0.04056 1.664 -0.012 0.147 fixed
count_birth_order3/5+ -0.009537 0.03866 -0.2467 -0.08532 0.06624 fixed
count_birth_order4/5+ -0.005476 0.03705 -0.1478 -0.07809 0.06714 fixed
count_birth_order5/5+ 0.03103 0.03971 0.7813 -0.0468 0.1089 fixed
count_birth_order5+/5+ 0.03933 0.03021 1.302 -0.01988 0.09853 fixed
sd_(Intercept).mother_pidlink 0 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2812 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_uterus_preg_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 575.8 636.8 -276.9 553.8 NA NA NA
12 577.8 644.3 -276.9 553.8 0.02339 1 0.8784
16 584.7 673.3 -276.3 552.7 1.123 4 0.8905
26 594.9 739 -271.4 542.9 9.793 10 0.4588

Parental full sibling order

outcome_parental_m1 <- update(m2_birthorder_linear, data = birthorder %>%
  mutate(sibling_count = sibling_count_genes_factor,
         birth_order_nonlinear = birthorder_genes_factor,
         birth_order = birthorder_genes,
         count_birth_order = count_birthorder_genes
         ) %>% 
      filter(sibling_count != "1"))
compare_models_markdown(outcome_parental_m1)
Basic Model
Model Summary
m1_covariates_only <- update(m2_birthorder_linear, formula = . ~ . - birth_order)
tidy(m1_covariates_only, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.04607 0.2413 -0.1909 -0.519 0.4269 fixed
poly(age, 3, raw = TRUE)1 0.0834 0.02552 3.267 0.03337 0.1334 fixed
poly(age, 3, raw = TRUE)2 -0.002838 0.0008763 -3.239 -0.004556 -0.001121 fixed
poly(age, 3, raw = TRUE)3 0.00002979 0.00000959 3.107 0.000011 0.00004859 fixed
male 0.2043 0.04246 4.81 0.121 0.2875 fixed
sibling_count3 0.002244 0.02053 0.1094 -0.03798 0.04247 fixed
sibling_count4 0.01715 0.02185 0.7848 -0.02568 0.05997 fixed
sibling_count5 0.01547 0.02478 0.6244 -0.0331 0.06405 fixed
sibling_count5+ 0.01676 0.02088 0.8027 -0.02417 0.05769 fixed
sd_(Intercept).mother_pidlink 0.0000001348 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2794 NA NA NA NA Residual
Coefficient Plot
plot(allEffects(m1_covariates_only))

Add Birth Order Linear
Model Summary
tidy(m2_birthorder_linear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.04579 0.2414 -0.1897 -0.5188 0.4272 fixed
birth_order -0.001402 0.004028 -0.3481 -0.009298 0.006493 fixed
poly(age, 3, raw = TRUE)1 0.08359 0.02554 3.273 0.03354 0.1336 fixed
poly(age, 3, raw = TRUE)2 -0.002842 0.0008765 -3.242 -0.00456 -0.001124 fixed
poly(age, 3, raw = TRUE)3 0.00002977 0.000009593 3.103 0.00001097 0.00004857 fixed
male 0.2039 0.04248 4.799 0.1206 0.2871 fixed
sibling_count3 0.002897 0.02062 0.1405 -0.03751 0.0433 fixed
sibling_count4 0.01871 0.02231 0.8385 -0.02502 0.06243 fixed
sibling_count5 0.01788 0.02573 0.6947 -0.03256 0.06831 fixed
sibling_count5+ 0.0218 0.02541 0.8578 -0.02801 0.07161 fixed
sd_(Intercept).mother_pidlink 0.0000001197 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2795 NA NA NA NA Residual
Coefficient Plot
plot_birthorder2(m2_birthorder_linear, separate = FALSE)

Add Birth Order Factor
Model Summary
m3_birthorder_nonlinear = update(m1_covariates_only, formula = . ~ . + birth_order_nonlinear)
tidy(m3_birthorder_nonlinear, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.04889 0.2419 -0.2021 -0.523 0.4252 fixed
poly(age, 3, raw = TRUE)1 0.08412 0.0256 3.286 0.03395 0.1343 fixed
poly(age, 3, raw = TRUE)2 -0.002866 0.0008787 -3.262 -0.004588 -0.001144 fixed
poly(age, 3, raw = TRUE)3 0.00003016 0.000009617 3.136 0.00001131 0.00004901 fixed
male 0.2047 0.04253 4.812 0.1213 0.288 fixed
sibling_count3 0.004311 0.0211 0.2044 -0.03704 0.04566 fixed
sibling_count4 0.01953 0.02331 0.8379 -0.02615 0.06521 fixed
sibling_count5 0.01357 0.02714 0.5 -0.03962 0.06676 fixed
sibling_count5+ 0.01085 0.02636 0.4114 -0.04083 0.06252 fixed
birth_order_nonlinear2 -0.01028 0.01744 -0.5894 -0.04446 0.0239 fixed
birth_order_nonlinear3 -0.01277 0.02068 -0.6177 -0.0533 0.02776 fixed
birth_order_nonlinear4 -0.003305 0.02593 -0.1275 -0.05412 0.04751 fixed
birth_order_nonlinear5 0.01257 0.03247 0.3873 -0.05106 0.0762 fixed
birth_order_nonlinear5+ 0.008002 0.03061 0.2614 -0.052 0.068 fixed
sd_(Intercept).mother_pidlink 0.0000001198 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2797 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m3_birthorder_nonlinear, separate = FALSE)

Add Interaction
Model Summary
m4_interaction = update(m3_birthorder_nonlinear, formula = . ~ . - birth_order_nonlinear - sibling_count + count_birth_order)
tidy(m4_interaction, conf.int = T)
term estimate std.error statistic conf.low conf.high group
(Intercept) -0.0656 0.2431 -0.2698 -0.5421 0.4109 fixed
poly(age, 3, raw = TRUE)1 0.08471 0.02575 3.29 0.03425 0.1352 fixed
poly(age, 3, raw = TRUE)2 -0.002891 0.000884 -3.27 -0.004623 -0.001158 fixed
poly(age, 3, raw = TRUE)3 0.00003046 0.000009677 3.148 0.0000115 0.00004943 fixed
male 0.2065 0.04261 4.845 0.1229 0.29 fixed
count_birth_order2/2 0.01718 0.03204 0.5362 -0.04562 0.07998 fixed
count_birth_order1/3 0.005085 0.02819 0.1804 -0.05016 0.06033 fixed
count_birth_order2/3 0.0003569 0.03145 0.01135 -0.06129 0.062 fixed
count_birth_order3/3 0.02669 0.03354 0.7957 -0.03905 0.09243 fixed
count_birth_order1/4 0.06555 0.0353 1.857 -0.003631 0.1347 fixed
count_birth_order2/4 -0.01125 0.03483 -0.323 -0.07952 0.05702 fixed
count_birth_order3/4 0.01287 0.03668 0.3507 -0.05903 0.08476 fixed
count_birth_order4/4 0.02968 0.03883 0.7645 -0.04641 0.1058 fixed
count_birth_order1/5 0.04971 0.0441 1.127 -0.03673 0.1361 fixed
count_birth_order2/5 0.02667 0.04994 0.534 -0.07122 0.1246 fixed
count_birth_order3/5 -0.01025 0.04789 -0.2141 -0.1041 0.08361 fixed
count_birth_order4/5 0.0333 0.04609 0.7225 -0.05703 0.1236 fixed
count_birth_order5/5 0.004253 0.04725 0.09001 -0.08835 0.09685 fixed
count_birth_order1/5+ 0.01471 0.04019 0.3659 -0.06407 0.09348 fixed
count_birth_order2/5+ 0.03404 0.04497 0.757 -0.0541 0.1222 fixed
count_birth_order3/5+ -0.01059 0.0404 -0.262 -0.08976 0.06859 fixed
count_birth_order4/5+ 0.007186 0.04081 0.1761 -0.07281 0.08718 fixed
count_birth_order5/5+ 0.05585 0.04037 1.384 -0.02326 0.135 fixed
count_birth_order5+/5+ 0.0298 0.02955 1.009 -0.02811 0.08771 fixed
sd_(Intercept).mother_pidlink 0.0000001199 NA NA NA NA mother_pidlink
sd_Observation.Residual 0.2799 NA NA NA NA Residual
Coefficient Plot
plot_birthorder(m4_interaction)

Model Comparison
###### Model 1 - Model 2
anova(m1_covariates_only, m2_birthorder_linear, m3_birthorder_nonlinear, m4_interaction)
## refitting model(s) with ML (instead of REML)
Data: birthorder %>% mutate(sibling_count = sibling_count_genes_factor, …
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
11 539.4 600.1 -258.7 517.4 NA NA NA
12 541.3 607.5 -258.7 517.3 0.1219 1 0.727
16 548.3 636.5 -258.2 516.3 0.9801 4 0.9128
26 561 704.3 -254.5 509 7.335 10 0.6935

Compare birth order specifications

library(coefplot)
multiplot(outcome_naive_m1, outcome_preg_m1, outcome_uterus_m1, outcome_parental_m1, dodgeHeight = 0.6,
          intercept = FALSE)

Multiplot Linear Birth Order Effect

Models

Intelligence = lmer(g_factor_2015_old ~ birthorder_genes + poly(age, 3, raw = TRUE) + male +
                        sibling_count_genes +
               (1 | mother_pidlink),
                   data = birthorder)
`Educational Attainment` = lmer(years_of_education_z ~ birthorder_genes + poly(age, 3, raw = TRUE) + male +
                                  sibling_count_genes + (1 | mother_pidlink),
                   data = birthorder)
Extraversion = lmer(big5_ext ~ birthorder_genes + poly(age, 3, raw = TRUE) + male +
                                  sibling_count_genes + (1 | mother_pidlink),
                   data = birthorder)
Neuroticism = lmer(big5_neu ~ birthorder_genes + poly(age, 3, raw = TRUE) + male +
                                  sibling_count_genes + (1 | mother_pidlink),
                   data = birthorder)
Conscientiousness = lmer(big5_con ~ birthorder_genes + poly(age, 3, raw = TRUE) + male +
                                  sibling_count_genes + (1 | mother_pidlink),
                   data = birthorder)
Agreeableness = lmer(big5_agree ~ birthorder_genes + poly(age, 3, raw = TRUE) + male +
                                  sibling_count_genes + (1 | mother_pidlink),
                   data = birthorder)
Openness = lmer(big5_open ~ birthorder_genes + poly(age, 3, raw = TRUE) + male +
                                  sibling_count_genes + (1 | mother_pidlink),
                   data = birthorder)
RiskA = lmer(riskA ~ birthorder_genes + poly(age, 3, raw = TRUE) + male +
                                  sibling_count_genes + (1 | mother_pidlink),
                   data = birthorder)
RiskB = lmer(riskB ~ birthorder_genes + poly(age, 3, raw = TRUE) + male +
                                  sibling_count_genes + (1 | mother_pidlink),
                   data = birthorder)

Plot

x = multiplot(Intelligence, `Educational Attainment`, Agreeableness, Conscientiousness, Extraversion, Neuroticism, Openness, RiskA, RiskB, intercept = FALSE, coefficients = c("birthorder_genes", "sibling_count_genes"), plot = F)
        
x = data.frame(x)

x$Model = ifelse(x$Model == "`Educational Attainment`", "Educational Attainment", x$Model)
x$Model = as.factor(x$Model)
levels(x$Model)

Agreeableness, Conscientiousness, Educational Attainment, Extraversion, Intelligence, Neuroticism, Openness, RiskA and RiskB

x$Coefficient = ifelse(x$Coefficient == "birthorder_genes", "Birth order", "Sibship size")

ggplot(x, aes(Model, Value)) +
  geom_pointrange(data=x, mapping=aes(x=Model, y=Value, ymin=HighOuter, ymax=LowOuter, color = Coefficient),
                  position = position_dodge(width=0.3), size = 1) +
  scale_colour_brewer(palette = "Set2") +
  geom_hline(yintercept = 0, size = 1.5)  +
  geom_hline(yintercept = -.1, linetype = "dotted", size = 1.5) + 
  scale_x_discrete(limits=c("RiskB", "RiskA", "Openness", "Neuroticism", "Extraversion",
                            "Conscientiousness", "Agreeableness", "Educational Attainment",
                            "Intelligence")) +
  scale_y_continuous(limits = c(-0.1, 0.05), breaks = c(-0.1, -0.05, 0, 0.05)) +
  coord_flip() +
  theme(text = element_text(size=25), axis.text.x = element_text(size = 20),
        axis.text.y = element_text(size = 20)) +
  labs(y = "Linear effect size", x = "Outcome")

---
title: " "
author: " "
output: html_document
editor_options: 
  chunk_output_type: console
---
# <span style="color:#E5C494">Birth Order Effects</span> {.tabset}

## Helper
```{r helper}
source("0_helpers.R")
```

## Load data
```{r Load Data}
birthorder = readRDS("data/alldata_birthorder.rds")
```

## Data preparations
```{r data preparations}
# For analyses we want to clean the dataset and get rid of all uninteresting data
birthorder = birthorder %>%
  filter(!is.na(pidlink)) %>% # no individuals who are only known from the pregnancy file
  filter(is.na(lifebirths) | lifebirths == 2) %>% # remove 7 and 2 individuals who are known as stillbirth or miscarriage but still have PID
  select(-lifebirths) %>%
  filter(!is.na(mother_pidlink)) %>%
  select(-father_pidlink) %>%
  filter(is.na(any_multiple_birth) | any_multiple_birth != 1) %>% # remove families with twins/triplets/..
  filter(!is.na(birthorder_naive)) %>%
  select(-starts_with("age_"), -wave, -any_multiple_birth, -multiple_birth) %>%
  mutate(money_spent_smoking_log = if_else(is.na(money_spent_smoking_log) & ever_smoked == 0, 0, money_spent_smoking_log),
         amount = if_else(is.na(amount) & ever_smoked == 0, 0, amount),
         amount_still_smokers = if_else(is.na(amount_still_smokers) &  still_smoking == 0, 0, amount_still_smokers),
         birthyear = lubridate::year(birthdate))

# recode Factor Variable as Dummy Variable
birthorder = left_join(birthorder,
                                birthorder %>%
                                  filter(!is.na(Category)) %>%
                                  mutate(var = 1) %>%
                                  select(pidlink, Category, var) %>%
                                  spread(Category, var, fill = 0, sep = "_"), by = "pidlink") %>%
  select(-Category)

# recode Factor Variable as Dummy Variable
birthorder = left_join(birthorder,
                                birthorder %>%
                                  filter(!is.na(Sector)) %>%
                                  mutate(var = 1) %>%
                                  select(pidlink, Sector, var) %>%
                                  spread(Sector, var, fill = 0, sep = "_"), by = "pidlink") %>%
  select(-Sector)

### Variables
birthorder = birthorder %>%
  mutate(
    # center variables that are used for analysis
  g_factor_2015_old = scale(g_factor_2015_old),
  g_factor_2015_young = scale(g_factor_2015_young),
  g_factor_2007_old = scale(g_factor_2007_old),
  g_factor_2007_young = scale(g_factor_2007_young),
  raven_2015_old = scale(raven_2015_old),
  math_2015_old = scale(math_2015_old),
  count_backwards = scale(count_backwards),
  raven_2015_young = scale(raven_2015_young),
  math_2015_young = scale(math_2015_young),
  words_remembered_avg = scale(words_remembered_avg),
  words_immediate = scale(words_immediate),
  words_delayed = scale(words_delayed),
  adaptive_numbering = scale(adaptive_numbering),
  raven_2007_old = scale(raven_2007_old),
  math_2007_old = scale(math_2007_old),
  raven_2007_young = scale(raven_2007_young),
  math_2007_young = scale(math_2007_young),
  riskA = scale(riskA),
  riskB = scale(riskB),
  years_of_education_z = scale(years_of_education),
  Total_score_highest_z = scale(Total_score_highest),
  wage_last_month_z = scale(wage_last_month_log),
  wage_last_year_z = scale(wage_last_year_log),
  big5_ext = scale(big5_ext),
  big5_con = scale(big5_con),
  big5_agree = scale(big5_agree),
  big5_open = scale(big5_open),
  big5_neu = scale(big5_neu),
  attended_school = as.integer(attended_school),
  attended_school = ifelse(attended_school == 1, 0,
                           ifelse(attended_school == 2, 1, NA)))


### Birthorder and Sibling Count
birthorder = birthorder %>% 
  mutate(
# birthorder as factors with levels of 1, 2, 3, 4, 5, 5+
    birthorder_naive_factor = as.character(birthorder_naive),
    birthorder_naive_factor = ifelse(birthorder_naive > 5, "5+",
                                            birthorder_naive_factor),
    birthorder_naive_factor = factor(birthorder_naive_factor, 
                                            levels = c("1","2","3","4","5","5+")),
    sibling_count_naive_factor = as.character(sibling_count_naive),
    sibling_count_naive_factor = ifelse(sibling_count_naive > 5, "5+",
                                               sibling_count_naive_factor),
    sibling_count_naive_factor = factor(sibling_count_naive_factor, 
                                               levels = c("2","3","4","5","5+")),

    birthorder_uterus_alive_factor = as.character(birthorder_uterus_alive),
    birthorder_uterus_alive_factor = ifelse(birthorder_uterus_alive > 5, "5+",
                                            birthorder_uterus_alive_factor),
    birthorder_uterus_alive_factor = factor(birthorder_uterus_alive_factor, 
                                            levels = c("1","2","3","4","5","5+")),
    sibling_count_uterus_alive_factor = as.character(sibling_count_uterus_alive),
    sibling_count_uterus_alive_factor = ifelse(sibling_count_uterus_alive > 5, "5+",
                                               sibling_count_uterus_alive_factor),
    sibling_count_uterus_alive_factor = factor(sibling_count_uterus_alive_factor, 
                                               levels = c("2","3","4","5","5+")),
    birthorder_uterus_preg_factor = as.character(birthorder_uterus_preg),
    birthorder_uterus_preg_factor = ifelse(birthorder_uterus_preg > 5, "5+",
                                           birthorder_uterus_preg_factor),
    birthorder_uterus_preg_factor = factor(birthorder_uterus_preg_factor,
                                           levels = c("1","2","3","4","5","5+")),
    sibling_count_uterus_preg_factor = as.character(sibling_count_uterus_preg),
    sibling_count_uterus_preg_factor = ifelse(sibling_count_uterus_preg > 5, "5+",
                                              sibling_count_uterus_preg_factor),
    sibling_count_uterus_preg_factor = factor(sibling_count_uterus_preg_factor, 
                                              levels = c("2","3","4","5","5+")),
    birthorder_genes_factor = as.character(birthorder_genes),
    birthorder_genes_factor = ifelse(birthorder_genes >5 , "5+", birthorder_genes_factor),
    birthorder_genes_factor = factor(birthorder_genes_factor, 
                                     levels = c("1","2","3","4","5","5+")),
    sibling_count_genes_factor = as.character(sibling_count_genes),
    sibling_count_genes_factor = ifelse(sibling_count_genes >5 , "5+",
                                        sibling_count_genes_factor),
    sibling_count_genes_factor = factor(sibling_count_genes_factor, 
                                        levels = c("2","3","4","5","5+")),
    # interaction birthorder * siblingcout for each birthorder
    count_birthorder_naive =
      factor(str_replace(as.character(interaction(birthorder_naive_factor,                                                              sibling_count_naive_factor)),
                        "\\.", "/"),
                                           levels =   c("1/2","2/2", "1/3",  "2/3",
                                                        "3/3", "1/4", "2/4", "3/4", "4/4",
                                                        "1/5", "2/5", "3/5", "4/5", "5/5",
                                                        "1/5+", "2/5+", "3/5+", "4/5+",
                                                        "5/5+", "5+/5+")),
    count_birthorder_uterus_alive =
      factor(str_replace(as.character(interaction(birthorder_uterus_alive_factor,                                                              sibling_count_uterus_alive_factor)),
                        "\\.", "/"),
                                           levels =   c("1/2","2/2", "1/3",  "2/3",
                                                        "3/3", "1/4", "2/4", "3/4", "4/4",
                                                        "1/5", "2/5", "3/5", "4/5", "5/5",
                                                        "1/5+", "2/5+", "3/5+", "4/5+",
                                                        "5/5+", "5+/5+")),
    count_birthorder_uterus_preg =
      factor(str_replace(as.character(interaction(birthorder_uterus_preg_factor,                                                              sibling_count_uterus_preg_factor)), 
                         "\\.", "/"),
                                           levels =   c("1/2","2/2", "1/3",  "2/3",
                                                        "3/3", "1/4", "2/4", "3/4", "4/4",
                                                        "1/5", "2/5", "3/5", "4/5", "5/5",
                                                        "1/5+", "2/5+", "3/5+", "4/5+",
                                                        "5/5+", "5+/5+")),
    count_birthorder_genes =
      factor(str_replace(as.character(interaction(birthorder_genes_factor,                                                              sibling_count_genes_factor)), "\\.", "/"),
                                           levels =   c("1/2","2/2", "1/3",  "2/3",
                                                        "3/3", "1/4", "2/4", "3/4", "4/4",
                                                        "1/5", "2/5", "3/5", "4/5", "5/5",
                                                        "1/5+", "2/5+", "3/5+", "4/5+",
                                                        "5/5+", "5+/5+")))

birthorder <- birthorder %>%
                     mutate(sibling_count = sibling_count_naive_factor,
                            birth_order_nonlinear = birthorder_naive_factor,
                            birth_order = birthorder_naive,
                            count_birth_order = count_birthorder_naive)

```


## Intelligence {.active .tabset}
### g-factor 2015 old {.tabset}

```{r g-factor 2015 old}
birthorder <- birthorder %>% mutate(outcome = g_factor_2015_old)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count +
               (1 | mother_pidlink),
                   data = birthorder)


compare_birthorder_specs(model)
```



### g-factor 2015 young {.tabset}

```{r g-factor 2015 young}
birthorder <- birthorder %>% mutate(outcome = g_factor_2015_young)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count +
               (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```


### g-factor 2007 old {.tabset}

```{r g-factor 2007 old}
birthorder <- birthorder %>% mutate(outcome = g_factor_2007_old)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```

### g-factor 2007 young {.tabset}

```{r g-factor 2007 young}
birthorder <- birthorder %>% mutate(outcome = g_factor_2007_young)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```


### Raven 2015 old {.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = raven_2015_old)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```

### Numeracy 2015 old {.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = math_2015_old)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```

### Raven 2015 young {.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = raven_2015_young)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```

### Numeracy 2015 young {.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = math_2015_young)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```

### Raven 2007 old {.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = raven_2007_old)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
             data = birthorder)
compare_birthorder_specs(model)
```

### Numeracy 2007 old{.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = math_2007_old)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
             data = birthorder)
compare_birthorder_specs(model)
```

### Raven 2007 young {.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = raven_2007_young)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
             data = birthorder)
compare_birthorder_specs(model)
```

### Numeracy 2007 young {.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = math_2007_young)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
             data = birthorder)
compare_birthorder_specs(model)
```

### Counting backwards {.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = count_backwards)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```

### Immediate word recall {.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = words_immediate)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```

### Delayed word recall {.tabset}
```{r}
birthorder <- birthorder %>% mutate(outcome = words_delayed)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```

### Adaptive Numbering {.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = adaptive_numbering)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```

## Personality {.tabset}

### Extraversion {.tabset}


```{r}
birthorder <- birthorder %>% mutate(outcome = big5_ext)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```


### Neuroticism {.tabset}


```{r}
birthorder <- birthorder %>% mutate(outcome = big5_neu)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```


### Conscientiousness {.tabset}


```{r}
birthorder <- birthorder %>% mutate(outcome = big5_con)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```


### Agreeableness {.tabset}


```{r}
birthorder <- birthorder %>% mutate(outcome = big5_agree)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```


### Openness {.tabset}


```{r}
birthorder <- birthorder %>% mutate(outcome = big5_open)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```


## Risk preference {.tabset}

### Risk A {.tabset}


```{r}
birthorder <- birthorder %>% mutate(outcome = riskA)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```


### Risk B {.tabset}


```{r}
birthorder <- birthorder %>% mutate(outcome = riskB)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```


## Educational Attainment {.tabset}
### Years of Education - z-standardized {.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = years_of_education_z)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```

### Elementary missed {.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = Elementary_missed)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```


### Elementary worked {.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = Elementary_worked)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```


### Attended School {.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = attended_school)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```


## Work {.tabset}

### Income Last Month (log) - z-standardized {.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = wage_last_month_z)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```

### Income last year (log) - z-standardized {.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = wage_last_year_z)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```

### Self-Employment - non standardized {.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = Self_employed)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```


## Work Category{.tabset}

### Category_Casual worker in agriculture{.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = `Category_Casual worker in agriculture`)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```

### Category_Casual worker not in agriculture{.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = `Category_Casual worker not in agriculture`)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```

### Category_Government worker{.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = `Category_Government worker`)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```


### Category_Private worker{.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = `Category_Private worker`)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```

### Category_Self-employed{.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = `Category_Self-employed`)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```

### Category_Unpaid family worker{.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = `Category_Unpaid family worker`)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```

## Work Sector {.tabset}

### Sector_Agriculture, forestry, fishing and hunting {.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = `Sector_Agriculture, forestry, fishing and hunting`)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```

### Sector_Construction {.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = `Sector_Construction`)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```

### Sector_Electricity, gas, water {.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = `Sector_Electricity, gas, water`)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```


### Sector_Finance, insurance, real estate and business services {.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = `Sector_Finance, insurance, real estate and business services`)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```


### Sector_Manufacturing {.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = Sector_Manufacturing)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```



### Sector_Mining and quarrying {.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = `Sector_Mining and quarrying`)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```


### Sector_Social services {.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = `Sector_Social services`)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```


### Sector_Transportation, storage and communications {.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = `Sector_Transportation, storage and communications`)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```

## Smoking Behaviour {.tabset}
### ever_smoked {.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = ever_smoked)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```

### still_smoking {.tabset}

```{r}
birthorder <- birthorder %>% mutate(outcome = still_smoking)
model = lmer(outcome ~ birth_order + poly(age, 3, raw = TRUE) + male + sibling_count + (1 | mother_pidlink),
                   data = birthorder)
compare_birthorder_specs(model)
```

## Multiplot Linear Birth Order Effect {.tabset}

### Models {.tabset}

```{r g-factor 2015 old}
Intelligence = lmer(g_factor_2015_old ~ birthorder_genes + poly(age, 3, raw = TRUE) + male +
                        sibling_count_genes +
               (1 | mother_pidlink),
                   data = birthorder)

```

```{r g-factor 2015 old}
`Educational Attainment` = lmer(years_of_education_z ~ birthorder_genes + poly(age, 3, raw = TRUE) + male +
                                  sibling_count_genes + (1 | mother_pidlink),
                   data = birthorder)


```



```{r g-factor 2015 old}
Extraversion = lmer(big5_ext ~ birthorder_genes + poly(age, 3, raw = TRUE) + male +
                                  sibling_count_genes + (1 | mother_pidlink),
                   data = birthorder)


```



```{r g-factor 2015 old}
Neuroticism = lmer(big5_neu ~ birthorder_genes + poly(age, 3, raw = TRUE) + male +
                                  sibling_count_genes + (1 | mother_pidlink),
                   data = birthorder)


```


```{r g-factor 2015 old}
Conscientiousness = lmer(big5_con ~ birthorder_genes + poly(age, 3, raw = TRUE) + male +
                                  sibling_count_genes + (1 | mother_pidlink),
                   data = birthorder)


```



```{r g-factor 2015 old}
Agreeableness = lmer(big5_agree ~ birthorder_genes + poly(age, 3, raw = TRUE) + male +
                                  sibling_count_genes + (1 | mother_pidlink),
                   data = birthorder)


```


```{r g-factor 2015 old}
Openness = lmer(big5_open ~ birthorder_genes + poly(age, 3, raw = TRUE) + male +
                                  sibling_count_genes + (1 | mother_pidlink),
                   data = birthorder)


```


```{r g-factor 2015 old}
RiskA = lmer(riskA ~ birthorder_genes + poly(age, 3, raw = TRUE) + male +
                                  sibling_count_genes + (1 | mother_pidlink),
                   data = birthorder)


```

```{r g-factor 2015 old}
RiskB = lmer(riskB ~ birthorder_genes + poly(age, 3, raw = TRUE) + male +
                                  sibling_count_genes + (1 | mother_pidlink),
                   data = birthorder)


```

### Plot {.active}
```{r}
x = multiplot(Intelligence, `Educational Attainment`, Agreeableness, Conscientiousness, Extraversion, Neuroticism, Openness, RiskA, RiskB, intercept = FALSE, coefficients = c("birthorder_genes", "sibling_count_genes"), plot = F)
        
x = data.frame(x)

x$Model = ifelse(x$Model == "`Educational Attainment`", "Educational Attainment", x$Model)
x$Model = as.factor(x$Model)
levels(x$Model)

x$Coefficient = ifelse(x$Coefficient == "birthorder_genes", "Birth order", "Sibship size")

ggplot(x, aes(Model, Value)) +
  geom_pointrange(data=x, mapping=aes(x=Model, y=Value, ymin=HighOuter, ymax=LowOuter, color = Coefficient),
                  position = position_dodge(width=0.3), size = 1) +
  scale_colour_brewer(palette = "Set2") +
  geom_hline(yintercept = 0, size = 1.5)  +
  geom_hline(yintercept = -.1, linetype = "dotted", size = 1.5) + 
  scale_x_discrete(limits=c("RiskB", "RiskA", "Openness", "Neuroticism", "Extraversion",
                            "Conscientiousness", "Agreeableness", "Educational Attainment",
                            "Intelligence")) +
  scale_y_continuous(limits = c(-0.1, 0.05), breaks = c(-0.1, -0.05, 0, 0.05)) +
  coord_flip() +
  theme(text = element_text(size=25), axis.text.x = element_text(size = 20),
        axis.text.y = element_text(size = 20)) +
  labs(y = "Linear effect size", x = "Outcome")
  
```
